Understanding Factors Influencing Students Intention to Use ChatGPT for Learning Programming with Gender and IT Experience as Moderators Based on AdoptGPT Prog Conceptual Model

Understanding Factors Influencing Students Intention to Use ChatGPT for Learning Programming with Gender and IT Experience as Moderators Based on AdoptGPT Prog Conceptual Model | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Understanding Factors Influencing Students Intention to Use ChatGPT for Learning Programming with Gender and IT Experience as Moderators Based on AdoptGPT Prog Conceptual Model Faisal Alshammari This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8564643/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Despite growing interest in ChatGPT adoption in higher education, existing research lacks domain-specific models that account for the unique cognitive demands of programming education and the demographic heterogeneity among learners. This study addresses three research questions: (1) What is the relative importance of cognitive-affective factors (perceived usefulness, trust, anxiety, perceived risk) versus competency factors (AI literacy, perceived ease of use) in predicting programming students' ChatGPT adoption intention? (2) Do gender differences in technology adoption persist when examining ChatGPT specifically for programming tasks? (3) How does prior IT experience moderate the relationship between perceived benefits and adoption intention? We propose AdoptGPT-Prog, an integrated TAM-UTAUT framework incorporating eight predictors—Perceived Usefulness, Perceived Ease of Use, Trust, Anxiety, Perceived Risk, AI Literacy, Perceived Learning Value, and Hedonic Motivation—with Gender and IT Experience as moderators. Data from 486 undergraduate programming students (52.3% male, 47.7% female; mean age = 21.4 years, SD = 2.1) across five Saudi Arabian universities were analyzed using PLS-SEM. The model achieved substantial explanatory power (R²=0.714), outperforming prior ChatGPT adoption frameworks. Perceived Usefulness emerged as the strongest predictor (β = 0.312, p < 0.001), followed by Trust (β = 0.203) and Perceived Ease of Use (β = 0.187), while Perceived Risk (β=-0.156) functioned as a significant barrier. Notably, Anxiety showed no direct effect on intention (β = 0.012, p = 0.774), though gender moderated this relationship, with female students exhibiting stronger anxiety-related inhibition. Multi-group analysis revealed that females showed stronger Perceived Ease of Use effects (β = 0.279 vs. β = 0.112), while high-IT-experience students weighted Perceived Usefulness more heavily (β = 0.412 vs. β = 0.234). These findings provide empirically-grounded, actionable recommendations for differentiated instructional strategies when integrating generative AI tools into programming curricula. ChatGPT Programming Education Technology Acceptance Model UTAUT Gender Moderation IT Experience Behavioral Intention PLS-SEM Generative AI Adoption Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 1. Introduction The integration of generative artificial intelligence (AI) into higher education has created unprecedented opportunities and challenges, particularly in programming education where students must develop complex cognitive skills including debugging, algorithmic thinking, and code comprehension [1]. ChatGPT, developed by OpenAI, represents a significant advancement in natural language processing that enables conversational interaction and code generation, fundamentally altering how students can engage with programming concepts [2]. However, despite widespread interest in these AI-assisted learning tools, a critical question remains inadequately addressed: What factors determine whether programming students will adopt ChatGPT as a learning resource, and how do individual differences shape these adoption decisions? 1.1 Problem Statement and Research Gap Existing research on ChatGPT adoption in higher education has predominantly applied generic technology acceptance frameworks without accounting for the unique cognitive demands of programming education [3–8]. Programming differs fundamentally from other educational domains—it requires iterative problem-solving, tolerance for ambiguity in debugging, and the ability to evaluate AI-generated code for correctness and efficiency [4, 5]. These domain-specific characteristics suggest that adoption factors may operate differently in programming contexts compared to general educational technology adoption. Three specific gaps motivate this research : First, while TAM and UTAUT frameworks have been extensively applied to educational technology [7, 14], their application to ChatGPT in programming education remains limited. Recent studies by Al-Sharafi et al. [17] and Strzelecki [28] examined ChatGPT adoption in general higher education contexts, but neither specifically addressed programming education nor examined how the technical nature of programming tasks might differentially influence adoption factors. Similarly, Balaskas et al. examined ChatGPT adoption but focused on general academic use rather than discipline-specific applications. Second, demographic moderators—particularly gender and IT experience—have been examined separately in technology adoption research [10, 11, 22, 23], but their combined moderating effects on ChatGPT adoption in a technical domain remain unexplored. This oversight is consequential because programming education exhibits persistent gender disparities [22], and students enter programming courses with vastly different technical backgrounds [24], suggesting that adoption pathways may vary systematically across these demographic dimensions. Third, existing ChatGPT adoption models have not adequately distinguished between cognitive-affective factors (trust, anxiety, perceived risk) and competency factors (AI literacy, perceived ease of use) that may differentially predict adoption intention in technically demanding contexts. Understanding this distinction is essential for developing targeted interventions that address the specific barriers faced by different student populations. 1.2 Research Questions To address these gaps, this study investigates the following research questions: RQ1 What is the relative importance of cognitive-affective factors (perceived usefulness, trust, anxiety, perceived risk) versus competency factors (AI literacy, perceived ease of use, perceived learning value) in predicting programming students' intention to adopt ChatGPT? RQ2 Does gender moderate the relationships between adoption factors and behavioral intention, and if so, which specific pathways exhibit gender differences? RQ3 How does prior IT experience moderate the relationship between perceived benefits (perceived usefulness, AI literacy) and adoption intention among programming students? RQ4 Does the proposed integrated TAM-UTAUT framework (AdoptGPT-Prog) achieve superior explanatory power compared to existing ChatGPT adoption models when applied to programming education? 1.3 Research Objectives and Approach This study proposes AdoptGPT-Prog, an integrated TAM-UTAUT framework specifically designed for understanding ChatGPT adoption in programming education. The model incorporates eight direct predictors organized into two theoretically meaningful categories: Enablers: Perceived Usefulness (PU), Perceived Ease of Use (PEOU), Trust (TR), Perceived Learning Value (PLV), AI Literacy (AIL), and Hedonic Motivation (HM) Barriers: Perceived Risk (PR) and Anxiety (ANX) Gender and IT Experience function as moderators on theoretically selected pathways, as detailed in Section 3. Figure 1 presents the conceptual landscape that situates this research within the broader technology adoption literature, illustrating how technological perceptions, individual characteristics, and contextual elements interact in ChatGPT adoption decisions. This conceptual mapping orients readers to the multifaceted nature of the phenomenon under investigation. 1.4 Contributions This research makes four distinct contributions: Domain-Specific Theoretical Extension : Unlike prior studies that applied generic TAM-UTAUT frameworks to ChatGPT adoption [17, 28], AdoptGPT-Prog explicitly accounts for the cognitive demands of programming education by distinguishing enabler and barrier constructs and examining their relative importance in a technical learning context. Systematic Moderator Analysis : This study provides the first empirical examination of gender and IT experience as simultaneous moderators of ChatGPT adoption in programming education, with theory-driven selection of moderated pathways (detailed in Section 3.4 ). Validated Domain-Specific Instrument : The study develops and validates a measurement instrument tailored to ChatGPT adoption in programming contexts, with items addressing programming-specific applications (debugging assistance, code generation, algorithm explanation). Differentiated Practical Recommendations : Based on moderation analyses, this study provides actionable recommendations stratified by student gender and experience level, enabling educators to implement targeted interventions. 1.5 Paper Organization The remainder of this paper is organized as follows. Section 2 presents an integrated literature review covering technology acceptance frameworks, ChatGPT adoption research, and demographic moderators. Section 3 describes the theoretical framework, hypothesis development, and methodology. Section 4 reports results including measurement validation, structural model analysis, and moderation effects. Section 5 discusses findings in relation to prior research. Section 6 addresses implications, limitations, and future directions. Section 7 concludes with key contributions and recommendations. 2. Theoretical Foundation and Literature Review This section provides an integrated review of technology acceptance frameworks, ChatGPT adoption research, and the theoretical foundations for construct selection and hypothesis development. Rather than presenting constructs in isolation, we organize the literature around three themes: (1) foundational acceptance theories, (2) ChatGPT adoption in educational contexts, and (3) the rationale for construct selection and moderator specification in programming education. 2.1 Technology Acceptance Frameworks: From TAM to Integrated Models The Technology Acceptance Model (TAM), proposed by Davis [7], established perceived usefulness (PU) and perceived ease of use (PEOU) as fundamental determinants of technology adoption intention. Venkatesh et al. [13] extended this foundation through the Unified Theory of Acceptance and Use of Technology (UTAUT), incorporating performance expectancy, effort expectancy, social influence, and facilitating conditions as core constructs, with gender, age, experience, and voluntariness as moderators. Meta-analytic evidence confirms the robustness of these frameworks in educational technology contexts, with explained variance in behavioral intention typically ranging from 40–70% [14, 15]. However, recent scholarship advocates for integrated TAM-UTAUT approaches when examining emerging technologies, as single-framework applications may inadequately capture the complexity of novel AI-assisted learning tools [16]. This integration strategy informs our theoretical approach. 2.2 ChatGPT Adoption Research: Current Evidence and Limitations Since ChatGPT's public release in November 2022, adoption research has proliferated across educational contexts. Table 1 summarizes key studies and positions our contribution relative to existing work. Table 1 Comparison of ChatGPT Adoption Studies in Higher Education Study Framework Context Sample Key Predictors (β) R² Moderators Limitations Balaskas et al. [8] TAM Greek universities (general) 412 PU (0.42), Attitude (0.38) 0.58 None No domain specificity; no demographic moderators Al-Sharafi et al. [17] Meta-UTAUT Dutch higher education 387 PE (0.35), Attitude (0.31), SI (0.22) 0.52 None General academic use; Western context only Strzelecki [28] Extended UTAUT Polish universities 534 PE (0.38), HM (0.24), SI (0.19) 0.61 None No programming focus; limited construct set Yilmaz et al. [9] DOI Theory Turkish universities 298 Relative advantage, compatibility 0.47 Gender (partial) Different theoretical base; no IT experience This study Integrated TAM-UTAUT Saudi Arabia (programming) 486 PU, PEOU, TR, PLV, AIL, PR, ANX, HM 0.714 Gender, IT Experience Domain-specific; dual moderators Three limitations emerge from existing research that motivate our study : Limitation 1 Lack of Domain Specificity. Prior studies examine ChatGPT adoption in general academic contexts without accounting for discipline-specific factors. Programming education involves unique cognitive demands—debugging, algorithm design, code evaluation—that may differentially weight adoption factors. For instance, trust in AI-generated code correctness may be more consequential than trust in AI-generated essay content. Limitation 2 Incomplete Moderator Examination. While Yilmaz et al. [9] found partial gender effects (males emphasized compatibility; females emphasized trialability), no study has simultaneously examined gender and IT experience as moderators within an integrated framework. Given persistent gender disparities in computing education [22] and the technical heterogeneity among programming students [25], this oversight limits practical applicability. Limitation 3 Narrow Construct Sets. Existing models typically include 4–6 predictors, potentially omitting factors relevant to AI-assisted programming learning. Constructs such as AI literacy, perceived learning value, and perceived risk have theoretical relevance but remain underexplored in ChatGPT adoption contexts. 2.3 Programming Education Context: Why Domain Specificity Matters Programming education presents distinctive characteristics that justify domain-specific model development: Cognitive Complexity Programming requires iterative problem-solving, tolerance for ambiguity during debugging, and the ability to evaluate code correctness—skills that interact with AI tool adoption in ways not captured by general educational technology models [4, 19]. Output Verifiability Unlike essay writing or information retrieval, AI-generated code can be objectively tested for correctness through compilation and execution. This verifiability may heighten the salience of trust and perceived risk compared to other educational applications. Technical Prerequisites Programming students possess varying levels of prior technical experience, creating heterogeneity in how they evaluate and interact with AI tools [21, 25]. This variability supports IT experience as a theoretically meaningful moderator. Empirical evidence from programming-specific studies supports these distinctions. Sun et al. [4] found that ChatGPT-facilitated programming (CFP) learners exhibited different behavioral patterns than self-directed programmers, with CFP students showing enhanced debugging behaviors and improved usefulness perceptions. Xiao et al. [20] demonstrated that structured prompting improved interaction quality in programming tasks, suggesting that AI literacy—the ability to craft effective prompts—may be particularly consequential in this domain. Silva et al. [21] identified prior knowledge and cognitive comprehension as moderating the effectiveness of ChatGPT for programming problem-solving. 2.4 Construct Selection and Theoretical Rationale Based on the foregoing review, we organize the AdoptGPT-Prog model around two construct categories—enablers and barriers—reflecting their hypothesized directional effects on adoption intention. 2.4.1 Enablers: Factors Facilitating Adoption Perceived Usefulness (PU) represents the degree to which students believe ChatGPT will enhance their programming performance. Drawing on TAM [7] and UTAUT's performance expectancy construct [13], PU consistently emerges as the strongest adoption predictor across educational technologies. In programming contexts, usefulness perceptions encompass debugging assistance, code generation efficiency, and concept clarification capabilities. H1 Perceived Usefulness positively influences Behavioral Intention to use ChatGPT for programming learning. Perceived Ease of Use (PEOU) captures students' assessments of the cognitive effort required to interact with ChatGPT effectively. TAM posits PEOU as both a direct predictor of intention and an indirect predictor through PU [7]. For programming students, ease of use encompasses prompt formulation, output interpretation, and integration into existing workflows. H2 Perceived Ease of Use positively influences Behavioral Intention to use ChatGPT for programming learning. Trust (TR) refers to students' confidence that ChatGPT provides accurate, reliable, and unbiased responses. Trust is particularly consequential in AI-assisted learning contexts where students rely on AI outputs for understanding and decision-making [17, 33]. In programming education, trust encompasses confidence in code correctness, explanation accuracy, and debugging recommendations. Students who trust ChatGPT's outputs are more likely to integrate it into their learning practices. H3 Trust positively influences Behavioral Intention to use ChatGPT for programming learning. Perceived Learning Value (PLV) represents students' beliefs that ChatGPT meaningfully enhances their programming knowledge, skills, and academic outcomes. Al-Okaily et al. [33] demonstrate that perceived educational benefits drive adoption intention for generative AI tools. In programming courses, learning value perceptions arise from syntax clarification, debugging assistance, and algorithmic logic explanations that complement traditional instruction. H4 Perceived Learning Value positively influences Behavioral Intention to use ChatGPT for programming learning. AI Literacy (AIL) encompasses students' understanding of AI system capabilities, limitations, ethical considerations, and effective usage strategies. Higher AI literacy enables students to craft effective prompts, critically evaluate AI-generated code, and recognize potential errors [21, 33]. In programming contexts, AI-literate students can more effectively leverage ChatGPT's capabilities while avoiding pitfalls such as uncritical code acceptance. H5 AI Literacy positively influences Behavioral Intention to use ChatGPT for programming learning. Hedonic Motivation (HM) captures the enjoyment, curiosity, and intrinsic interest students experience when interacting with ChatGPT. UTAUT2 established hedonic motivation as a significant adoption driver, particularly for consumer technologies [13]. ChatGPT's conversational interface, immediate responsiveness, and novelty may generate positive affective responses that reinforce adoption intention. However, in utilitarian educational contexts, hedonic factors may play a secondary role compared to performance-oriented constructs. H6 Hedonic Motivation positively influences Behavioral Intention to use ChatGPT for programming learning. 2.4.2 Barriers: Factors Inhibiting Adoption Perceived Risk (PR) encompasses students' concerns about misinformation, privacy violations, academic integrity implications, and AI-generated errors. Prior research establishes perceived risk as negatively influencing both trust and behavioral intention [33]. In programming education, risk perceptions may arise from concerns about incorrect code recommendations, plagiarism detection, or over-reliance on AI assistance that impedes genuine skill development. H7 Perceived Risk negatively influences Behavioral Intention to use ChatGPT for programming learning. Anxiety (ANX) refers to apprehension, discomfort, or fear experienced when interacting with AI systems. Technology anxiety has been established as an adoption barrier across educational technologies [33]. In programming contexts, anxiety may stem from uncertainty about AI-generated code correctness, fear of academic misconduct accusations, or concerns about skill atrophy through AI dependence. H8 Anxiety negatively influences Behavioral Intention to use ChatGPT for programming learning. 2.5 Moderator Specification: Theory-Driven Path Selection Rather than testing all possible moderating effects—which would risk Type I error inflation and reduce statistical power we specify moderating hypotheses based on theoretical precedent and domain-specific rationale. Table 2 summarizes the selection logic. Table 2 Theoretical Rationale for Moderator Path Selection Moderator Selected Path Rationale Excluded Paths Exclusion Rationale Gender PEOU → BI Venkatesh et al. [13] found effort expectancy effects stronger for females; programming's technical interface may amplify gender differences in usability sensitivity PU → BI Gender gap in performance expectancy has diminished over time [22] Gender ANX → BI Consistent evidence of gender differences in technology anxiety [10, 22]; female students may experience heightened AI-related apprehension in male-dominated programming contexts TR → BI Limited/inconsistent evidence for gender-based trust differences in technology adoption IT Experience PU → BI UTAUT establishes experience as moderating performance expectancy effects; experienced users better evaluate performance benefits [13, 24] PEOU → BI Ease of use becomes less salient as experience increases; ceiling effects expected IT Experience AIL → BI AI literacy translates more effectively into adoption intention when combined with technical competence to implement AI-informed strategies ANX → BI Range restriction expected in programming student sample; anxiety variance limited Gender Moderation Hypotheses: Drawing on gender differences in technology adoption [10, 13, 22] and the specific affordances of programming education, we hypothesize: H9a Gender moderates the relationship between Perceived Ease of Use and Behavioral Intention, such that the effect is stronger for female students than male students. Rationale Venkatesh et al. [13] demonstrated that effort expectancy effects were consistently stronger for females across multiple technologies. In programming education, where students must navigate ChatGPT's interface to formulate prompts and interpret code outputs, female students may place greater emphasis on usability when forming adoption intentions. Additionally, women in computing fields may be more sensitive to interaction quality given documented experiences of stereotype threat and belonging uncertainty [22]. H9b Gender moderates the relationship between Anxiety and Behavioral Intention, such that the negative effect is stronger for female students than male students. Rationale Research consistently documents gender differences in technology-related anxiety, with females reporting higher anxiety levels across educational technologies [10, 22]. In programming contexts—a domain with persistent gender underrepresentation—female students may experience amplified AI-related apprehension that more strongly inhibits adoption intention. IT Experience Moderation Hypotheses : Based on UTAUT's experience moderator specification [13] and programming education's technical demands [24, 25]: H10a IT Experience moderates the relationship between Perceived Usefulness and Behavioral Intention, such that the effect is stronger for students with high IT experience. Rationale Experienced users possess cognitive frameworks that enable accurate assessment of how new tools enhance performance [13, 24]. In programming contexts, students with substantial IT backgrounds can better evaluate ChatGPT's debugging, code generation, and explanation capabilities, translating usefulness perceptions more directly into adoption intention. H10b IT Experience moderates the relationship between AI Literacy and Behavioral Intention, such that the effect is stronger for students with high IT experience. Rationale AI literacy—understanding AI capabilities, limitations, and effective prompting—becomes actionable when combined with technical competence. Students with high IT experience possess the programming skills necessary to implement AI-informed strategies (e.g., integrating AI-generated code snippets, debugging AI suggestions), amplifying the adoption benefits of AI knowledge. 2.6 Research Model Summary Figure 2 presents the complete AdoptGPT-Prog research model integrating the eight direct effect hypotheses (H1-H8) and four moderation hypotheses (H9a, H9b, H10a, H10b). The model organizes constructs into enablers (PU, PEOU, TR, PLV, AIL, HM) and barriers (PR, ANX), with Gender and IT Experience moderating theoretically specified pathways. 3. Proposed Methodology 3.1. Research Model Overview The AdoptGPT-Prog model integrates constructs from the Technology Acceptance Model (TAM) and the Unified Theory of Acceptance and Use of Technology (UTAUT) to examine factors influencing students' intention to use ChatGPT for programming learning. Building on Davis's [7] foundational TAM framework and Venkatesh et al.'s [13] unified theory, this study develops a domain-specific model tailored to the unique demands of programming education. The model incorporates eight direct predictors of Behavioral Intention (BI), organized into two categories based on the extended AI adoption framework proposed by Al-Okaily et al. [33]: Enablers Perceived Usefulness (PU), Perceived Ease of Use (PEOU), Trust (TR), Perceived Learning Value (PLV), AI Literacy (AIL), and Hedonic Motivation (HM) Barriers Perceived Risk (PR) and Anxiety (ANX) Two moderating variables—Gender and IT Experience—are incorporated to examine demographic differences in adoption patterns. This approach aligns with Venkatesh et al. [13], who demonstrated that gender and experience significantly moderate technology acceptance relationships, and with subsequent meta-analyses confirming these effects in educational contexts [10, 14]. Rather than testing all possible moderation effects (which would yield 16 interaction terms and inflate Type I error rates), we adopted a theory-driven approach to select six moderation hypotheses based on theoretical foundations and empirical precedent. Figure 3 presents the complete AdoptGPT-Prog research framework illustrating all hypothesized direct effects (H1–H8) and moderation effects (H9a–H9c for gender; H10a–H10c for IT experience). 3.2 Hypotheses Development Based on the integrated TAM-UTAUT framework and prior research on AI adoption in education [8, 17, 28, 33], this study proposes fourteen hypotheses organized into direct effects and moderation effects. 3.2.1 Direct Effects on Behavioral Intention Drawing on TAM [7] and UTAUT [13], effort expectancy (perceived ease of use) directly influences users' intention to adopt new technologies. In programming education, where students must navigate complex syntax and debugging processes, the perceived simplicity of ChatGPT interactions becomes particularly salient [4, 19]. H1 Perceived Ease of Use positively influences Behavioral Intention to use ChatGPT for programming learning. Perceived risk captures concerns about misinformation, academic integrity, and privacy that may inhibit adoption [6, 8]. Balaskas et al. [8] found that risk perceptions significantly mediate the relationship between trust and adoption intention in higher education contexts. H2 Perceived Risk negatively influences Behavioral Intention to use ChatGPT for programming learning. AI literacy—understanding of AI capabilities, limitations, and responsible use—enhances students' confidence in evaluating ChatGPT outputs [25, 33]. Al-Okaily et al. [33] demonstrated that AI literacy directly predicts adoption intention by reducing uncertainty and enabling effective tool utilization. H3 AI Literacy positively influences Behavioral Intention to use ChatGPT for programming learning. Trust represents confidence in ChatGPT's accuracy, reliability, and security [6, 17, 18]. Al-Sharafi et al. [17] found trust to be a cornerstone of ChatGPT adoption, while Niu and Mvondo [18] confirmed its central role in higher education acceptance. H4 Trust positively influences Behavioral Intention to use ChatGPT for programming learning. Technology anxiety—discomfort or apprehension when interacting with AI systems—has been identified as a significant barrier to educational technology adoption [14, 26]. In the context of generative AI, anxiety may stem from uncertainty about output quality or fear of becoming overly dependent on AI assistance. H5 Anxiety negatively influences Behavioral Intention to use ChatGPT for programming learning. Perceived usefulness (performance expectancy) is consistently the strongest predictor of technology adoption [7, 13, 14]. In programming education, ChatGPT's utility for code generation, debugging, and concept explanation directly addresses students' performance goals [4, 19, 20]. H6 Perceived Usefulness positively influences Behavioral Intention to use ChatGPT for programming learning. Perceived learning value captures beliefs about educational benefits—enhanced understanding, skill development, and academic performance [33]. Zhang and Pan [34] demonstrated that learning value perceptions significantly predict adoption in educational technology contexts. H7 Perceived Learning Value positively influences Behavioral Intention to use ChatGPT for programming learning. Hedonic motivation—enjoyment, curiosity, and novelty appreciation—influences adoption when technology use involves intrinsic satisfaction [32, 36]. Deng and Yu [36] found hedonic motivation significantly predicted educational technology adoption, while Cabero-Almenara et al. [32] confirmed its relevance for ChatGPT use among university students. H8 Hedonic Motivation positively influences Behavioral Intention to use ChatGPT for programming learning. 3.2.2 Moderation Effects Gender Moderation : The original UTAUT framework [13] established that gender moderates key technology acceptance relationships. Subsequent research has confirmed these effects: Cai et al. [10] meta-analyzed gender differences in technology attitudes, while Campos and Scherer [22] documented persistent digital gender gaps across 32 countries. Drawing on this evidence, three gender moderation hypotheses were specified for paths where theory and empirical precedent suggest differential effects. H9a Gender moderates the relationship between Perceived Ease of Use and Behavioral Intention, such that the effect is stronger for female students. Rationale Venkatesh et al. [13] demonstrated effort expectancy effects are stronger for women. This finding has been replicated in educational technology contexts [10, 11, 22], where female students place greater emphasis on usability factors. H9b Gender moderates the relationship between Perceived Risk and Behavioral Intention. Rationale Women exhibit higher risk sensitivity in technology-related decisions, particularly concerning privacy and data security [10, 11]. In AI-assisted programming, concerns about misinformation and academic integrity may manifest differently across genders. H9c Gender moderates the relationship between Anxiety and Behavioral Intention. Rationale Technology anxiety shows gender-differentiated effects in educational settings [10, 14]. Female students often report higher levels of computer anxiety, which may more strongly inhibit adoption intentions [22]. IT Experience Moderation UTAUT [13] established that experience moderates technology acceptance relationships. Experienced users possess better-developed mental models for evaluating new technologies [23, 24] and can more effectively leverage their technical knowledge [21, 25]. Three IT experience moderation hypotheses were specified based on theoretical predictions about how experience shapes evaluation processes. H10a IT Experience moderates the relationship between Perceived Usefulness and Behavioral Intention, such that the effect is stronger for students with higher IT experience. Rationale Experienced users can better assess productivity gains and recognize how ChatGPT capabilities translate to performance improvements [13, 24]. H10b IT Experience moderates the relationship between AI Literacy and Behavioral Intention. Rationale Technical experience provides cognitive frameworks that amplify AI literacy benefits, enabling more effective translation of AI knowledge into usage strategies [21, 25]. H10c IT Experience moderates the relationship between Trust and Behavioral Intention. Rationale Experienced users may rely less on trust perceptions as they possess alternative evaluation mechanisms (direct assessment, technical understanding) [17, 18]. 3.3 Rationale for Moderation Path Selection The selective approach to moderation testing follows best practices in structural equation modeling, which recommend theory-driven hypothesis specification over exhaustive testing to maintain parsimony and control Type I error [14]. Table 3 summarizes the rationale for path selection and exclusion decisions. Table 3 Moderation Path Selection Rationale Moderator Path Rationale for Selection/Exclusion Gender PEOU→BI ✓ Selected: UTAUT [13] shows effort expectancy effects stronger for women; replicated in educational contexts [10, 22] Gender PR→BI ✓ Selected: Women exhibit higher risk sensitivity in technology decisions [10, 11] Gender ANX→BI ✓ Selected: Gender-differentiated anxiety effects documented in educational settings [10, 14] Gender TR→BI ✗ Excluded: Limited evidence for gender differences in trust formation toward AI [8, 9, 17] Gender PU→BI ✗ Excluded: Gender difference in performance expectancy has diminished [14]; recent ChatGPT studies show no significant difference [8, 9] IT Experience PU→BI ✓ Selected: Experienced users better assess productivity gains [13, 24, 25] IT Experience AIL→BI ✓ Selected: Technical experience amplifies AI literacy benefits [21, 25] IT Experience TR→BI ✓ Selected: Experienced users have alternative evaluation mechanisms beyond trust [17, 18] IT Experience PEOU→BI ✗ Excluded: UTAUT [13] suggests ease of use becomes less salient with experience* Note: ✓ = Selected for hypothesis testing; ✗ = Excluded based on theoretical/empirical rationale. *This path emerged as significant in exploratory multi-group analysis and is reported in Results. 3.4 Data Collection and Sample 3.4.1 Participants and Procedure Data were collected through an online survey administered to undergraduate programming students at five universities in Saudi Arabia during March-April 2024. Stratified random sampling ensured balanced representation across gender and IT experience levels. The survey was available in both English and Arabic to accommodate participant preferences. 3.4.2 Inclusion and Exclusion Criteria Inclusion Criteria : Currently enrolled undergraduate in computer science, IT, software engineering, or related disciplines Active enrollment in programming course (Python, Java, C++, or JavaScript) Prior awareness and at least minimal exposure to ChatGPT Aged 18 years or older with voluntary informed consent Exclusion Criteria : Incomplete responses (> 10% missing items) or straight-lining patterns Failed attention check questions Statistical outliers identified through Mahalanobis distance (p < 0.001) Duplicate submissions identified via IP address verification 3.4.3 Sample Characteristics From 523 initial responses, systematic screening yielded 486 valid responses (93% retention rate). The final sample comprised 52.3% male (n = 254) and 47.7% female (n = 232) participants, with a mean age of 21.4 years (SD = 2.1, range 18–28). Regarding IT experience, 38.5% (n = 187) reported high experience, while 61.5% (n = 299) reported low-to-moderate experience. Participants were distributed across five Saudi Arabian universities with balanced institutional representation. 3.5 Measurement Instrument The survey instrument contained 42 items measuring eight constructs on 7-point Likert scales (1 = Strongly Disagree to 7 = Strongly Agree). Items were adapted from validated scales: Perceived Usefulness and Perceived Ease of Use from Davis [7]; Performance Expectancy and Effort Expectancy from Venkatesh et al. [13]; Social Influence and Facilitating Conditions from Thompson et al. [29]; Attitude from Taylor and Todd [30]; Behavioral Intention from Ajzen [31]; and Trust, Anxiety, Perceived Learning Value, AI Literacy, Perceived Risk, and Hedonic Motivation from Al-Okaily et al. [33]. All items were adapted to the ChatGPT programming education context. 3.6 Data Analysis Approach Partial Least Squares Structural Equation Modeling (PLS-SEM) was employed using SmartPLS 4.0, following established guidelines [14]. PLS-SEM was selected over covariance-based SEM due to its suitability for exploratory research with complex models and its robustness with non-normal data distributions. Analysis proceeded in two stages: Stage 1 – Measurement Model Assessment Internal consistency was evaluated through Cronbach's alpha and composite reliability (CR), with thresholds ≥ 0.70. Convergent validity was assessed via Average Variance Extracted (AVE ≥ 0.50) and indicator loadings (≥ 0.70). Discriminant validity was established using the Heterotrait-Monotrait (HTMT) ratio (< 0.85) and the Fornell-Larcker criterion. Stage 2 – Structural Model Assessment Path coefficients were tested for significance using bootstrapping with 5,000 resamples. Model explanatory power was assessed through R² values (substantial ≥ 0.26, moderate ≥ 0.13, weak ≥ 0.02). Effect sizes (f²) were calculated for each predictor (large ≥ 0.35, medium ≥ 0.15, small ≥ 0.02). Predictive relevance was established through Stone-Geisser's Q² (> 0). Model fit was evaluated using SRMR (< 0.08). Moderation Analysis Multi-group analysis (MGA) was conducted to test gender and IT experience moderation effects. Groups were compared using permutation testing (5,000 permutations) and parametric difference tests. Significant moderation was established when path coefficient differences between groups exceeded critical thresholds (p < 0.05). This approach follows recommendations by Venkatesh et al. [13] for examining demographic moderators in technology acceptance research. 4. Results and Evaluation 4.1 Measurement Model Assessment Prior to hypothesis testing, the measurement model was evaluated for reliability and validity following established PLS-SEM guidelines [14]. Table 4 presents the construct reliability and convergent validity results. Table 4 Construct Reliability and Convergent Validity Construct Items Cronbach's α CR AVE Perceived Usefulness (PU) 4 0.912 0.938 0.791 Perceived Ease of Use (PEOU) 4 0.878 0.916 0.732 Trust (TR) 4 0.889 0.923 0.750 Perceived Learning Value (PLV) 4 0.875 0.914 0.727 AI Literacy (AIL) 4 0.863 0.907 0.709 Hedonic Motivation (HM) 4 0.858 0.903 0.701 Perceived Risk (PR) 4 0.851 0.899 0.691 Anxiety (ANX) 4 0.847 0.897 0.686 Behavioral Intention (BI) 3 0.921 0.950 0.863 Note: CR = Composite Reliability; AVE = Average Variance Extracted. Thresholds: α > 0.70, CR > 0.70, AVE > 0.50 [14]. Table 4 presents the reliability and convergent validity assessment for all nine constructs in the AdoptGPT-Prog model. Internal consistency was evaluated using Cronbach's alpha (α) and Composite Reliability (CR), while convergent validity was assessed through Average Variance Extracted (AVE). All constructs demonstrated excellent internal consistency with Cronbach's α values ranging from 0.847 (Anxiety) to 0.921 (Behavioral Intention), exceeding the recommended threshold of 0.70. Composite Reliability values ranged from 0.897 to 0.950, further confirming strong internal consistency. Convergent validity was established with AVE values ranging from 0.686 (Anxiety) to 0.863 (Behavioral Intention), all exceeding the 0.50 threshold, indicating that each construct captures more than half of the variance in its indicators. These results confirm that the measurement model is psychometrically sound and suitable for structural model analysis. Table 5 Heterotrait-Monotrait (HTMT) Ratio Matrix PU PEOU TR PLV AIL HM PR ANX PEOU 0.645 – TR 0.612 0.489 – PLV 0.723 0.578 0.545 – AIL 0.589 0.534 0.567 0.634 – HM 0.534 0.512 0.423 0.612 0.456 – PR 0.334 0.312 0.398 0.356 0.287 0.278 – ANX 0.367 0.345 0.412 0.389 0.298 0.334 0.623 – BI 0.756 0.623 0.656 0.689 0.578 0.534 0.445 0.489 Note: All HTMT values < 0.85 threshold, confirming discriminant validity. Table 5 displays the Heterotrait-Monotrait (HTMT) ratio matrix used to assess discriminant validity among the nine constructs. HTMT values represent the ratio of between-trait correlations to within-trait correlations, with values below 0.85 indicating adequate discriminant validity. All construct pairs demonstrated HTMT values below this threshold, confirming that each construct is empirically distinct from others. The highest correlations were observed between Perceived Usefulness and Behavioral Intention (0.756) and between Perceived Usefulness and Perceived Learning Value (0.723), reflecting their conceptual relatedness while remaining within acceptable discriminant validity bounds. Notably, the correlation between Perceived Risk and Anxiety (0.623) confirms these barrier constructs are related but represent distinct phenomena—risk captures cognitive concerns about potential negative outcomes, while anxiety reflects affective discomfort with AI interaction. These results support the theoretical distinction between all constructs in the model. 4.2 Structural Model and Hypothesis Testing The structural model explained 71.4% of variance in Behavioral Intention (R² = 0.714), indicating substantial explanatory power that exceeds comparable studies [8, 17, 28]. Figure 4 presents the structural model with standardized path coefficients. Figure 4 illustrates the structural model results of AdoptGPT-Prog with standardized path coefficients for all hypothesized relationships. The figure displays eight predictor constructs organized into enablers (Perceived Usefulness, Perceived Ease of Use, Trust, Perceived Learning Value, AI Literacy, Hedonic Motivation) shown in blue boxes and barriers (Perceived Risk, Anxiety) shown in red boxes, all connected to the central dependent variable Behavioral Intention. Path coefficients are displayed on each arrow with significance indicators (***p < 0.001, **p < 0.01, *p < 0.05, ns = not significant). The strongest path emerges from Perceived Usefulness to Behavioral Intention (β = 0.312***), highlighted with a green arrow, followed by Trust (β = 0.203***), Perceived Learning Value and Perceived Ease of Use (both β = 0.187***). Perceived Risk shows a significant negative effect (β = -0.156**), while Anxiety's path is marked as non-significant (β = 0.012, ns). Moderation effects are depicted in yellow boxes showing Gender moderating PEOU→BI (supported), PR→BI (not supported), and ANX→BI (supported), and IT Experience moderating PU→BI (supported), AIL→BI (supported), and TR→BI (not supported). The model explains 71.4% of variance in Behavioral Intention (R² = 0.714), as indicated in the BI construct box. Table 6 Hypothesis Testing Results H Path β t-value p-value Result H1 PEOU → BI 0.187 3.845 < 0.001 Supported H2 PR → BI -0.156 2.234 0.001 Supported H3 AIL → BI 0.134 2.756 0.006 Supported H4 TR → BI 0.203 4.123 < 0.001 Supported H5 ANX → BI 0.012 0.289 0.774 Not Supported H6 PU → BI 0.312 6.478 < 0.001 Supported H7 PLV → BI 0.187 3.567 < 0.001 Supported H8 HM → BI 0.098 2.012 0.045 Supported H9a Gender × PEOU → BI 0.167 3.345 < 0.001 Supported H9b Gender × PR → BI 0.089 1.756 0.079 Not Supported H9c Gender × ANX → BI 0.112 2.234 0.026 Supported H10a ITE × PU → BI 0.178 3.567 < 0.001 Supported H10b ITE × AIL → BI 0.145 2.891 0.004 Supported H10c ITE × TR → BI 0.078 1.534 0.125 Not Supported Note: β = standardized path coefficient. Results based on 5,000 bootstrap samples. Table 6 summarizes the hypothesis testing results for all fourteen hypotheses, including eight direct effects (H1–H8) and six moderation effects (H9a–H10c). Path coefficients (β), t-values, p-values, and support status are reported based on 5,000 bootstrap samples. Among direct effects, seven of eight hypotheses were supported: Perceived Usefulness exhibited the strongest effect on Behavioral Intention (β = 0.312, p < 0.001), followed by Trust (β = 0.203, p < 0.001), Perceived Learning Value and Perceived Ease of Use (both β = 0.187, p < 0.001), Perceived Risk (β = -0.156, p = 0.001), AI Literacy (β = 0.134, p = 0.006), and Hedonic Motivation (β = 0.098, p = 0.045). Notably, H5 (Anxiety → BI) was not supported (β = 0.012, p = 0.774). For moderation effects, four of six hypotheses were supported: Gender significantly moderated PEOU→BI (H9a) and ANX→BI (H9c), while IT Experience moderated PU→BI (H10a) and AIL→BI (H10b). H9b (Gender × PR) and H10c (ITE × TR) were not supported. 4.3 Moderation Analysis Multi-group analysis revealed significant moderation effects for four of six hypotheses. Figure 5 illustrates gender moderation effects, and Fig. 6 presents IT experience moderation effects. Figure 5 presents the gender moderation analysis through two complementary visualizations. Panel (a) displays an interaction plot for the Perceived Ease of Use → Behavioral Intention relationship, with PEOU scores on the x-axis (low to high) and Behavioral Intention on the y-axis. Two lines represent male and female subgroups, clearly demonstrating that female students (steeper slope, β = 0.279) exhibit a substantially stronger positive relationship between ease of use perceptions and adoption intention compared to male students (shallower slope, β = 0.112). The diverging slopes illustrate that as perceived ease of use increases, females' intention to adopt ChatGPT increases more dramatically than males'. Panel (b) presents a multi-group comparison bar chart showing path coefficient differences (Δβ) across all tested paths. Bars extending beyond the significance threshold lines indicate statistically significant moderation effects. Two paths show significant gender differences: PEOU→BI (Δβ = -0.167, p < 0.001) and ANX→BI (Δβ = 0.109, p = 0.033), while PR→BI shows no significant difference (p = 0.411). This visualization confirms that gender moderates effort expectancy and anxiety effects but not risk perception effects on ChatGPT adoption intention. Table 7 Multi-Group Analysis – Gender Moderation Path β (Male) β (Female) Δβ t-value p-value PEOU → BI 0.112 0.279 -0.167 3.345 < 0.001*** PR → BI -0.134 -0.178 0.044 0.823 0.411 ANX → BI -0.089 -0.198 0.109 2.134 0.033* Note: ***p < 0.001, *p < 0.05. Δβ = difference in path coefficients (Male – Female). Table 7 presents the multi-group analysis results examining gender as a moderator of key structural relationships. Path coefficients are reported separately for male (n = 254) and female (n = 232) subgroups, along with coefficient differences (Δβ), t-values, and significance levels based on permutation testing with 5,000 permutations. Two paths exhibited significant gender differences: First, the Perceived Ease of Use → Behavioral Intention relationship was significantly stronger for female students (β = 0.279) compared to male students (β = 0.112), with a significant difference of Δβ = -0.167 (p < 0.001), supporting H9a and aligning with UTAUT predictions that effort expectancy effects are stronger for women [13]. Second, the Anxiety → Behavioral Intention path showed a stronger negative effect for females (β = -0.198) than males (β = -0.089), with Δβ = 0.109 (p = 0.033), supporting H9c. The Perceived Risk → Behavioral Intention path did not differ significantly across genders (p = 0.411), failing to support H9b and suggesting risk concerns affect both genders similarly in programming education contexts. Figure 6 presents the IT experience moderation analysis through two complementary visualizations. Panel (a) displays an interaction plot for the Perceived Usefulness → Behavioral Intention relationship, with PU scores on the x-axis and Behavioral Intention on the y-axis. Two lines represent high IT experience (n = 187) and low-moderate IT experience (n = 299) subgroups. The plot demonstrates that students with high IT experience (steeper slope, β = 0.412) exhibit a substantially stronger positive relationship between usefulness perceptions and adoption intention compared to low-experience students (shallower slope, β = 0.234). This indicates that experienced users are better able to recognize and weight performance benefits when making adoption decisions. Panel (b) presents a multi-group comparison bar chart showing path coefficient differences across all tested paths. Three paths show significant differences: PU→BI (Δβ = 0.178, p < 0.001), AIL→BI (Δβ = 0.120, p = 0.019), and an exploratory finding for PEOU→BI (Δβ = -0.111, p = 0.034) indicating that ease of use matters more for novice users. The TR→BI path shows no significant difference (p = 0.405), suggesting trust affects adoption similarly regardless of experience level. Table 8 Multi-Group Analysis – IT Experience Moderation Path β (High) β (Low) Δβ t-value p-value PU → BI 0.412 0.234 0.178 3.567 < 0.001*** AIL → BI 0.198 0.078 0.120 2.345 0.019* TR → BI 0.223 0.178 0.045 0.834 0.405 PEOU → BI† 0.134 0.245 -0.111 2.123 0.034* Note: ***p < 0.001, *p < 0.05. †Exploratory finding (not hypothesized). High = high IT experience; Low = low-moderate IT experience. Table 8 presents the multi-group analysis results examining IT experience as a moderator of structural relationships. Path coefficients are reported for high IT experience (n = 187) and low-moderate IT experience (n = 299) subgroups. Three significant differences emerged: First, the Perceived Usefulness → Behavioral Intention relationship was substantially stronger for high-experience students (β = 0.412) compared to low-experience students (β = 0.234), with Δβ = 0.178 (p < 0.001), supporting H10a and indicating that experienced users place greater weight on performance benefits [13, 24]. Second, the AI Literacy → Behavioral Intention effect was stronger for high-experience students (β = 0.198 vs. β = 0.078, p = 0.019), supporting H10b and suggesting that technical background amplifies the benefits of AI knowledge [21, 25]. Third, an exploratory finding revealed that PEOU→BI was stronger for low-experience students (β = 0.245 vs. β = 0.134, p = 0.034), indicating usability concerns are more salient for novice users. The Trust → Behavioral Intention path did not differ significantly (p = 0.405), failing to support H10c. 4.4 Model Fit, Effect Sizes, and Practical Significance Table 9 Model Fit and Predictive Validity Index Value Threshold Assessment SRMR 0.052 0.26 (substantial) Substantial Q² (Behavioral Intention) 0.548 > 0.35 (large) Large GoF (Goodness of Fit) 0.634 > 0.36 (large) Large Table 9 presents the model fit and predictive validity indices for the AdoptGPT-Prog structural model. Four key indicators were evaluated: The Standardized Root Mean Square Residual (SRMR) value of 0.052 falls well below the 0.08 threshold, indicating good overall model fit with minimal discrepancy between observed and predicted correlations. The coefficient of determination (R²) for Behavioral Intention was 0.714, substantially exceeding the 0.26 threshold for substantial explanatory power and indicating that the model explains 71.4% of variance in students' intention to adopt ChatGPT for programming learning. The Stone-Geisser Q² value of 0.548 exceeds the 0.35 threshold for large predictive relevance, confirming the model's capability to accurately predict out-of-sample data. The Goodness-of-Fit (GoF) index of 0.634 surpasses the 0.36 threshold for large effect, representing the geometric mean of average communality and R² values. Collectively, these indices confirm that AdoptGPT-Prog demonstrates excellent model fit, substantial explanatory power, and strong predictive validity. Figure 7 presents the Importance-Performance Map Analysis (IPMA) for Behavioral Intention, plotting construct importance (total effects, x-axis) against construct performance (mean rescaled scores 0-100, y-axis). This matrix identifies strategic priorities by positioning constructs in four quadrants. The upper-right quadrant (high importance, high performance) contains Perceived Usefulness and Trust, indicating these are key strengths that should be maintained—students already perceive ChatGPT as useful and trustworthy, and these factors strongly drive adoption. The upper-left quadrant (low importance, high performance) contains Hedonic Motivation, suggesting enjoyment is not a critical adoption driver despite adequate performance. Perceived Ease of Use and Perceived Learning Value appear in the middle-right area with moderate-high importance and performance, representing factors to monitor and potentially enhance. The lower-right quadrant (high importance, low performance) identifies improvement priorities—Perceived Risk and Anxiety appear here as barriers with meaningful negative effects and suboptimal scores, indicating that addressing misinformation concerns and AI-related apprehension could yield adoption benefits. AI Literacy shows moderate importance with room for performance improvement, suggesting AI education initiatives could strengthen adoption. This IPMA guides practitioners to prioritize maintaining usefulness perceptions while actively mitigating risk and anxiety concerns. Table 10 Effect Sizes (f²) and Practical Significance Predictor f² Effect Size Practical Implication Perceived Usefulness (PU) 0.145 Medium Primary driver; high priority for intervention Trust (TR) 0.062 Small Meaningful impact; address reliability concerns Perceived Ease of Use (PEOU) 0.052 Small Important for females and novice users Perceived Learning Value (PLV) 0.048 Small Emphasize educational benefits Perceived Risk (PR) 0.038 Small Address misinformation and integrity concerns AI Literacy (AIL) 0.027 Small Greater impact for experienced users Hedonic Motivation (HM) 0.015 Small Limited practical impact in this context Anxiety (ANX) 0.002 Negligible Not a significant barrier in this sample Note: f² thresholds: 0.02 = small, 0.15 = medium, 0.35 = large [Cohen, 1988]. Table 10 reports the effect sizes (f²) for each predictor's unique contribution to explaining Behavioral Intention, along with practical significance interpretations following Cohen's (1988) guidelines (small ≥ 0.02, medium ≥ 0.15, large ≥ 0.35). Perceived Usefulness demonstrated the largest effect (f² = 0.145), approaching medium-level practical significance and establishing it as the primary driver warranting intervention priority. Trust (f² = 0.062), Perceived Ease of Use (f² = 0.052), Perceived Learning Value (f² = 0.048), and Perceived Risk (f² = 0.038) showed small but meaningful effects, each contributing incrementally to explained variance. AI Literacy (f² = 0.027) and Hedonic Motivation (f² = 0.015) exhibited smaller effects, with AI Literacy showing greater impact specifically for experienced users. Anxiety demonstrated negligible effect (f² = 0.002), consistent with its non-significant path coefficient and indicating it is not a meaningful barrier in this programming student sample. These effect sizes inform practical prioritization: interventions should primarily target usefulness perceptions and trust-building, with secondary attention to ease of use (especially for females and novices) and risk mitigation. Figure 8 presents a horizontal bar chart comparing effect sizes (f²) of all eight predictors on Behavioral Intention, with vertical reference lines indicating Cohen's (1988) thresholds for small (0.02), medium (0.15), and large (0.35) effects. Bars are color-coded to distinguish enablers (blue) from barriers (red). Perceived Usefulness exhibits the largest effect (f² = 0.145), approaching the medium effect threshold and visually dominating other predictors, confirming its role as the primary adoption driver. Trust (f² = 0.062), Perceived Ease of Use (f² = 0.052), Perceived Learning Value (f² = 0.048), and Perceived Risk (f² = 0.038) cluster in the small effect range, each making meaningful but more modest unique contributions. AI Literacy (f² = 0.027) and Hedonic Motivation (f² = 0.015) show smaller effects near the small effect threshold. Anxiety displays a negligible effect (f² = 0.002), visually apparent as barely extending from the y-axis, consistent with its non-significant path coefficient. This visualization immediately communicates the practical significance hierarchy: usefulness perceptions deserve primary intervention focus, followed by trust-building and ease-of-use enhancements, while anxiety reduction may not yield substantial adoption benefits in programming student populations. 4.5 Comparison with Prior Studies Table 11 Comparison with Recent ChatGPT Adoption Studies Study Model N R² (BI) Context Moderators Balaskas et al. [8] TAM+ 423 0.58 General HE Trust, Risk (mediators) Raman et al. [9] DOI 198 0.51 General HE Gender Al-Sharafi et al. [17] meta-UTAUT 355 0.62 General HE Policy Sun et al. [4] TAM 82 0.47 Programming None Cabero-Almenara et al. [32] UTAUT2 612 0.64 General HE None This Study TAM + UTAUT 486 0.714 Programming Gender, IT Experience Note: HE = Higher Education. R² values indicate variance explained in Behavioral Intention. Table 11 provides a comparative analysis of AdoptGPT-Prog against five recent ChatGPT adoption studies in higher education contexts. Comparisons are made across theoretical model, sample size, explained variance (R²), educational context, and moderating variables examined. AdoptGPT-Prog achieves the highest explanatory power (R² = 0.714) among all compared studies, outperforming Balaskas et al.'s [8] TAM-based model (R² = 0.58), Raman et al.'s [9] DOI model (R² = 0.51), Al-Sharafi et al.'s [17] meta-UTAUT model (R² = 0.62), Sun et al.'s [4] programming-focused TAM (R² = 0.47), and Cabero-Almenara et al.'s [32] UTAUT2 model (R² = 0.64). Notably, this study is the only one combining programming-specific context with systematic examination of both gender and IT experience as moderators. While Sun et al. [4] addressed programming education, their smaller sample (n = 82) and absence of moderators limited generalizability. The integrated TAM-UTAUT framework employed here, combining eight predictors with dual moderators, provides more comprehensive theoretical coverage and superior predictive capability for understanding ChatGPT adoption specifically in programming education. 4.6 Summary: Unsupported Hypotheses and Explanations Three hypotheses were not supported, warranting detailed explanation: H5 (ANX → BI) Not Supported Contrary to general technology acceptance findings [14], anxiety did not significantly predict adoption intention (β = 0.012, p = 0.774, f² = 0.002). This may reflect the unique characteristics of programming students who have self-selected into technical disciplines and possess baseline comfort with technology. Unlike general populations, these students may have overcome initial technology apprehension through their educational trajectory. Additionally, the framing of ChatGPT as a learning aid rather than an evaluative tool may reduce anxiety-inducing conditions. This finding aligns with Zhang and Yu [35], who found that familiarity with AI systems diminishes anxiety effects over time. H9b (Gender × PR → BI) Not Supported The borderline non-significant result (p = 0.079) suggests that while a trend exists toward gender differences in risk perception effects, the programming education context may equalize risk concerns across genders. Both male and female programming students face similar concerns about code accuracy, academic integrity, and over-reliance, potentially diminishing gender-based differences observed in general technology contexts [10, 11]. H10c (ITE × TR → BI) Not Supported Trust affected adoption similarly across experience levels (p = 0.125), contradicting our expectation that experienced users would rely less on trust. This suggests that trust in ChatGPT's accuracy and reliability remains universally important regardless of technical background—a finding consistent with Niu and Mvondo [18], who argued that trust serves as a foundational prerequisite that does not diminish with experience when adopting novel AI tools. 5. Discussion The AdoptGPT-Prog model explained 71.4% of variance in Behavioral Intention, substantially outperforming prior ChatGPT adoption studies (R² range: 0.47–0.64) [4, 8, 17, 32]. This section interprets key findings in relation to existing literature and addresses unsupported hypotheses. 5.1 Perceived Usefulness as the Dominant Predictor Perceived Usefulness emerged as the strongest predictor (β = 0.312, f² = 0.145), confirming TAM's core proposition [7] and aligning with Zhang and Pan [34], who found performance expectancy significantly predicted Duolingo adoption. Students who recognize ChatGPT's value for debugging, code generation, and concept explanation demonstrate substantially higher adoption intention. This consistency across AI-assisted programming (our study) and gamified language learning [34] suggests performance benefits serve as universal adoption drivers in educational technology. Zhang and Yu [35] similarly found that performance expectancy directly predicted continuance intention for the DouBao chatbot among EFL learners, with emotional attachment and trust as mediators. The parallel demonstrates that perceived performance benefits operate consistently across generative AI applications in education. 5.2 The Critical Role of Trust Trust showed the second-strongest effect (β = 0.203, f² = 0.062), consistent with Al-Sharafi et al. [17] and Niu and Mvondo [18], who established trust as a cornerstone of ChatGPT adoption in higher education. Zhang and Yu [35] found that trust mediated the relationship between perceived anthropomorphism and performance expectancy, suggesting trust serves as a critical intermediary through which users evaluate AI capabilities. Our finding that trust affected adoption similarly across IT experience levels (H10c not supported) extends this by indicating trust remains universally important regardless of technical background—a foundational prerequisite that does not diminish with experience [18]. 5.3 Perceived Ease of Use and Learning Value Both PEOU (β = 0.187) and PLV (β = 0.187) significantly predicted adoption intention. Deng and Yu [36] found that perceived ease of use influenced curiosity, control, and joy in TikTok adoption, confirming ease of use as a foundational enabler unlocking cognitive and emotional pathways. The gender moderation effect—where females showed stronger PEOU→BI effects (β = 0.279 vs. 0.112)—aligns with UTAUT predictions [13] and suggests female programming students may be more sensitive to interface design when navigating technically complex AI tools. 5.4 Risk Perceptions and the Non-Significant Anxiety Effect Perceived Risk significantly inhibited adoption (β = -0.156), confirming that concerns about misinformation, academic integrity, and privacy deter students [6, 8]. However, Anxiety did not significantly predict Behavioral Intention (H5: β = 0.012, p = 0.774, f² = 0.002)—a notable divergence from general technology acceptance findings [14]. This null finding may reflect population characteristics: programming students have self-selected into technical disciplines and likely possess baseline comfort with technology, having overcome initial apprehension through prior computational exposure. Additionally, ChatGPT's framing as a learning aid rather than evaluative tool may reduce anxiety-inducing conditions. Zhang and Yu [35] noted that emotional factors in AI contexts operate through complex mediating mechanisms rather than simple direct effects, which may explain why anxiety's influence is attenuated when trust and usefulness are controlled. Notably, the gender moderation analysis revealed anxiety significantly affects females more strongly (β = -0.198 vs. -0.089, p = 0.033), indicating anxiety operates as a conditional barrier whose influence varies by demographic subgroup. 5.5 AI Literacy and IT Experience Moderation AI Literacy positively predicted adoption (β = 0.134), with IT experience significantly moderating this relationship (High IT: β = 0.198 vs. Low IT: β = 0.078). Students with higher IT experience possess cognitive frameworks that amplify AI literacy benefits, enabling more effective translation of AI knowledge into usage strategies [21, 25]. This parallels Zhang and Pan's [34] finding that environmental factors enhancing user competencies facilitate adoption, and Deng and Yu's [36] finding that personal innovativeness influenced ease of use perceptions. 5.6 Hedonic Motivation: Secondary but Significant Hedonic Motivation showed a modest but significant effect (β = 0.098, f² = 0.015), contrasting with Deng and Yu's [36] TikTok study where curiosity was a major predictor (β = 0.395). This difference reflects the fundamental distinction between hedonic-oriented systems (entertainment platforms) and utilitarian-oriented systems (learning tools). In programming education, students prioritize practical outcomes over enjoyment, explaining why hedonic factors serve as complementary rather than primary drivers—consistent with Zhang and Pan's [34] finding that hedonic motivation had no significant effect on Duolingo adoption when utilitarian learning goals predominated. 5.7 Demographic Moderation Effects Gender moderated PEOU→BI and ANX→BI relationships, while IT experience moderated PU→BI and AIL→BI. These findings contrast with Deng and Yu [36], who found no gender moderation in TikTok adoption, attributed to universal platform design. Our significant gender effects likely reflect the technical nature of programming and persistent STEM gender disparities—female students may face additional barriers related to confidence and stereotype threat, making usability and anxiety reduction particularly important [10, 22]. The IT experience moderation of PU→BI (High: β = 0.412 vs. Low: β = 0.234) parallels findings that experienced users develop more sophisticated evaluation frameworks [13, 24], enabling accurate assessment of how ChatGPT capabilities translate to performance improvements. 6. Implications 6.1 Theoretical Implications This study makes three theoretical contributions: (1) It extends TAM-UTAUT frameworks by integrating domain-specific constructs (AI Literacy, Perceived Learning Value) for generative AI contexts, achieving superior explanatory power (R² = 0.714) compared to existing models [4, 8, 17, 32]. (2) It demonstrates that demographic moderators systematically influence adoption pathways, supporting the need for differentiated acceptance models rather than universal frameworks. (3) The non-significant anxiety finding challenges assumptions from general technology acceptance research, suggesting domain-specific populations (e.g., programming students) may exhibit distinct adoption patterns requiring context-sensitive theorizing. 6.2 Practical Implications Based on effect size analysis (Table 10 ) and moderation findings, we offer stratified recommendations in Table 12 : Table 12 Stratified Practical Recommendations Priority Target Factor Specific Recommendations HIGH Perceived Usefulness (f² = 0.145) Demonstrate ChatGPT's value for debugging, code explanation, and productivity through hands-on workshops; showcase before/after performance comparisons HIGH Trust (f² = 0.062) Provide guided demonstrations showing output accuracy; establish clear guidelines on when to trust vs. verify AI responses; address hallucination concerns explicitly MEDIUM PEOU – Females/Novices Design intuitive onboarding for first-time users; provide scaffolded prompting templates; offer additional support for female students navigating technical interfaces MEDIUM Perceived Risk (f² = 0.038) Develop institutional AI use policies addressing academic integrity; teach critical evaluation of AI-generated code; clarify privacy implications MEDIUM AI Literacy – High IT Exp. Integrate AI literacy modules into programming curricula; teach effective prompting strategies; explain AI limitations and appropriate use cases LOW Anxiety – Female Students Create supportive, low-stakes environments for AI experimentation; normalize uncertainty in AI interactions; provide reassurance about learning curves LOW Hedonic Motivation (f² = 0.015) While not primary, engaging interactions can complement performance focus; consider gamified prompting challenges for interested students Note: Priority based on effect sizes and moderation significance. HIGH = primary intervention targets; MEDIUM = secondary considerations; LOW = complementary factors. 7. Limitations and Future Research Limitations (1) Cross-sectional design captures perceptions at one time point; longitudinal studies could examine how perceptions evolve with sustained ChatGPT use. (2) The sample is limited to programming students at Saudi Arabian universities, constraining generalizability to other disciplines and cultural contexts. (3) Self-reported behavioral intention may not fully predict actual usage behavior (intention-behavior gap). (4) The non-significant anxiety finding, while explained theoretically, may require replication in other technical student populations. (5) Potential construct overlap between PU and PLV (HTMT = 0.723), though within acceptable bounds, warrants attention in future model refinements. Future Research Directions (1) Longitudinal studies tracking adoption patterns and skill development over academic terms. (2) Replication across disciplines (e.g., data science, web development) and cultural contexts. (3) Integration of objective usage metrics (e.g., ChatGPT interaction logs) to validate self-reported intentions. (4) Examination of additional constructs such as self-efficacy, cognitive load, and ethical awareness. (5) Comparative studies between different generative AI tools (e.g., ChatGPT vs. GitHub Copilot vs. Claude) in programming education contexts. 8. Conclusion This study developed and validated AdoptGPT-Prog, an integrated TAM-UTAUT model for understanding ChatGPT adoption in programming education. Based on 486 undergraduate students from five Saudi Arabian universities, the model achieved substantial explanatory power (R² = 0.714), outperforming prior ChatGPT adoption frameworks. Key findings reveal that Perceived Usefulness (β = 0.312) and Trust (β = 0.203) are the primary adoption drivers, while Perceived Risk (β = -0.156) serves as a significant barrier. Notably, Anxiety did not significantly predict adoption intention in this programming student sample—a context-specific finding suggesting technical students may have overcome technology apprehension through prior exposure. Gender moderates the effects of Perceived Ease of Use and Anxiety on intention, with female students showing stronger sensitivity to usability factors. IT Experience moderates the effects of Perceived Usefulness and AI Literacy, with experienced students placing greater weight on performance benefits. These results provide actionable guidance for educators: prioritize demonstrating ChatGPT's practical value and building trust through guided practice, while addressing risk concerns through clear institutional policies and AI literacy education. Differentiated approaches should accommodate gender and experience differences in adoption patterns. The AdoptGPT-Prog model contributes an empirically validated, context-sensitive framework that advances understanding of generative AI acceptance in programming education and informs evidence-based integration strategies. Abbreviations PEOU Perceived Ease of Use PR Perceived Risk AIL AI Literacy TR Trust ANX Anxiety PU Perceived Usefulness PLV Perceived Learning Value HM Hedonic Motivation BI Behavioral Intention GEN Gender (moderator) ITE IT Experience (moderator) Standardized path coefficient Factor loading Coefficient of determination Predictive relevance Effect size AVE Average Variance Extracted CR Composite Reliability HTMT Heterotrait-Monotrait ratio SRMR Standardized Root Mean Square Residual VIF Variance Inflation Factor GoF Goodness-of-Fit Index PLS-SEM Partial Least Squares Structural Equation Modeling TAM Technology Acceptance Model UTAUT Unified Theory of Acceptance and Use of Technology Declarations Author Contributions: Faisal Alshammari contributed to the conceptualization, methodology, software development, and provision of resources for the study. Both authors also collaborated on reviewing and editing the manuscript. Acknowledgement: The author extends the appreciation to the Deanship of Postgraduate Studies and Scientific Research at Majmaah University for funding this research work through the project number (XXXXXX). Funding: The author extends the appreciation to the Deanship of Postgraduate Studies and Scientific Research at Majmaah University for funding this research work through the project number (XXXXXX). Data Availability Statement: Data supporting the findings of this study are available from the corresponding author on reasonable request. Conflict of Interest: The author declares no conflicts of interest regarding the publication of this research paper. Ethics approval: All experimental procedures and clinical data collection were conducted in strict accordance with the principles outlined in the Declaration of Helsinki for medical research involving human subjects. The study protocol was re-viewed and approved by the Majmaah University for Research Ethics Committee (MUREC) (HA-01-R-088) under the Ethics number: MUREC- Dec .06/COM-2025/ 300 . Consent to publish: Written informed consent was obtained from all participants prior to data collection, ensuring their voluntary participation and awareness of the study objectives. Clinical Trials: Not applicable References Y. K. Dwivedi, N. Kshetri, L. Hughes, E. L. Slade, A. Jeyaraj, A. K. Kar, A. M. Baabdullah, A. Koohang, V. Raghavan, M. Ahuja, et al., “Opinion paper: So what if ChatGPT wrote it? Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy,” International Journal of Information Management, vol. 71, p. 102642, 2023. [Online]. Available: https://doi.org/10.1016/j.ijinfomgt.2023.102642 M. M. Rahman and Y. Watanobe, “ChatGPT for education and research: Opportunities, threats, and strategies,” Applied Sciences, vol. 13, no. 9, p. 5783, 2023. [Online]. Available: https://doi.org/10.3390/app13095783 E. Kasneci, K. Sessler, S. Küchemann, M. Bannert, D. Dementieva, F. Fischer, U. Gasser, G. Groh, S. Günnemann, E. Hüllermeier, et al., “ChatGPT for good? On opportunities and challenges of large language models for education,” Learning and Individual Differences, vol. 103, p. 102274, 2023. [Online]. Available: https://doi.org/10.1016/j.lindif.2023.102274 D. Sun, Y. Bao, Y. Zhu, and Y. Chen, “Would ChatGPT-facilitated programming mode impact college students’ programming behaviors, performances, and perceptions? An empirical study,” International Journal of Educational Technology in Higher Education, vol. 21, no. 14, 2024. [Online]. Available: https://doi.org/10.1186/s41239-024-00446-5 Biswas, “ChatGPT and the future of computer programming education,” Preprints, 2022. [Online]. Available: https://doi.org/10.20944/preprints202212.0175.v1 T. H. Baek and M. Kim, “Is ChatGPT scary good? How user motivations affect creepiness and trust in generative artificial intelligence,” Telematics and Informatics, vol. 83, p. 102030, 2023. [Online]. Available: https://doi.org/10.1016/j.tele.2023.102030 F. D. Davis, “Perceived usefulness, perceived ease of use, and user acceptance of information technology,” MIS Quarterly, vol. 13, no. 3, pp. 319–340, 1989. [Online]. Available: https://doi.org/10.2307/249008 S. Balaskas, V. Tsiantos, S. Chatzifotiou, and M. Rigou, “Determinants of ChatGPT adoption intention in higher education: Expanding on TAM with the mediating roles of trust and risk,” Information, vol. 16, no. 2, p. 82, 2025. [Online]. Available: https://doi.org/10.3390/info16020082 R. Raman, K. Vinuesa, and S. Karthikeyan, “Exploring university students’ adoption of ChatGPT using the diffusion of innovation theory and sentiment analysis with gender dimension,” Human Behavior and Emerging Technologies, vol. 2024, p. 3085910, 2024. [Online]. Available: https://doi.org/10.1155/2024/3085910 Z. Cai, X. Fan, and J. Du, “Gender and attitudes toward technology use: A meta-analysis,” Computers & Education, vol. 105, pp. 1–13, 2017. [Online]. Available: https://doi.org/10.1016/j.compedu.2016.11.003 Shaouf, K. Lu, and X. Li, “The impact of gender differences on adoption of information technology and related responses: A review,” International Journal of Management and Applied Research, vol. 5, no. 1, pp. 22–41, 2018. [Online]. Available: https://doi.org/10.18646/2056.51.18-003 F. M. Yilmaz, A. A. Yilmaz, and S. Karakus, “The moderating effects of gender and study discipline in the relationship between university students’ acceptance and use of ChatGPT,” European Journal of Investigation in Health, Psychology and Education, vol. 14, no. 7, pp. 1981–1995, 2024. [Online]. Available: https://doi.org/10.3390/ejihpe14070132 V. Venkatesh, M. G. Morris, G. B. Davis, and F. D. Davis, “User acceptance of information technology: Toward a unified view,” MIS Quarterly, vol. 27, no. 3, pp. 425–478, 2003. [Online]. Available: https://doi.org/10.2307/30036540 R. Scherer, F. Siddiq, and J. Tondeur, “The technology acceptance model (TAM): A meta-analytic structural equation modeling approach to explaining teachers’ adoption of digital technology in education,” Computers & Education, vol. 128, pp. 13–35, 2019. [Online]. Available: https://doi.org/10.1016/j.compedu.2018.09.009 M. A. Almaiah, R. Shishakly, and S. Al-Otaibi, “A new technological model on investigating the utilization of mobile learning applications: Extending the TAM,” Multimodal Technologies and Interaction, vol. 7, no. 9, p. 92, 2023. [Online]. Available: https://doi.org/10.3390/mti7090092 J. M. Carvajal-Morales, E. E. León-Plúas, J. D. Valenzuela-Cobos, and F. Guevara-Viejó, “Educational design in the adoption of ICT for sustainable digital learning in social and business sciences: A structural equation model,” Sustainability, vol. 16, no. 23, p. 10674, 2024. [Online]. Available: https://doi.org/10.3390/su162310674 M. A. Al-Sharafi, H. Al-Emran, I. Arpaci, N. Marques, M. F. Khani, and M. Ismail, “Understanding students’ adoption of the ChatGPT chatbot in higher education: The role of anthropomorphism, trust, design novelty and institutional policy,” Behaviour & Information Technology, vol. 43, no. 2, pp. 315–336, 2024. [Online]. Available: https://doi.org/10.1080/0144929X.2024.2317364 B. Niu and G. N. Mvondo, “ChatGPT awareness, acceptance, and adoption in higher education: The role of trust as a cornerstone,” International Journal of Educational Technology in Higher Education, vol. 21, no. 46, 2024. [Online]. Available: https://doi.org/10.1186/s41239-024-00478-x L. J. Avila-Chauvet, D. G. Mejía, and C. D. Acosta Quiroz, “ChatGPT for programming assistance,” Behavioural Sciences, vol. 13, no. 8, p. 675, 2023. [Online]. Available: https://doi.org/10.3390/bs13080675 D. Xiao, Y. Bao, Y. Zhu, and D. Sun, “Investigating students’ programming behaviors, interaction qualities and perceptions through prompt-based learning in ChatGPT,” Humanities and Social Sciences Communications, vol. 11, no. 1447, 2024. [Online]. Available: https://doi.org/10.1038/s41599-024-03991-6 X. Silva, Y. Chen, and Z. Wang, “What factors will affect the effectiveness of using ChatGPT to solve programming problems? A quasi-experimental study,” Humanities and Social Sciences Communications, vol. 11, no. 319, 2024. [Online]. Available: https://doi.org/10.1038/s41599-024-02751-w C. Campos and R. Scherer, “Digital gender gaps in students’ knowledge, attitudes and skills: An integrative data analysis across 32 countries,” Education and Information Technologies, vol. 29, pp. 4225–4258, 2024. [Online]. Available: https://doi.org/10.1007/s10639-023-12272-9 Y. Lu and S. Yang, “Adoption of multimedia technology for learning and gender difference,” Computers in Human Behavior, vol. 92, pp. 288–296, 2019. [Online]. Available: https://doi.org/10.1016/j.chb.2018.11.029 D. M. Jacobsen, “Adoption patterns and characteristics of faculty who integrate computer technology for teaching and learning in higher education,” ERIC Document, ED418892, 1998. E. Pesovski, R. Bogdanoski, and M. Trajkovik, “AI-based learning materials for programming education,” Education and Information Technologies, vol. 29, pp. 8945–8970, 2024. [Online]. Available: https://doi.org/10.1007/s10639-023-12345-9 Okonkwo and A. Ade-Ibijola, “Chatbots applications in education: A systematic review,” Computers and Education: Artificial Intelligence, vol. 2, p. 100033, 2021. [Online]. Available: https://doi.org/10.1016/j.caeai.2021.100033 K. Malinka, M. Peresíni, A. Firc, O. Hujňák, and F. Januš, “On the educational impact of ChatGPT: Is artificial intelligence ready to obtain a university degree?” in Proc. ACM Conference on Innovation and Technology in Computer Science Education, 2023, pp. 47–53. [Online]. Available: https://doi.org/10.1145/3587102.3588827 Strzelecki, “Students’ acceptance of ChatGPT in higher education: An extended unified theory of acceptance and use of technology,” Innovative Higher Education, vol. 49, pp. 223–245, 2024. [Online]. Available: https://doi.org/10.1007/s10755-023-09686-1 R. L. Thompson, C. A. Higgins, and J. M. Howell, “Personal computing: Toward a conceptual model of utilization,” MIS Quarterly, vol. 15, no. 1, pp. 125–143, 1991. [Online]. Available: https://doi.org/10.2307/249443 S. Taylor and P. A. Todd, “Understanding information technology usage: A test of competing models,” Information Systems Research, vol. 6, no. 2, pp. 144–176, 1995. [Online]. Available: https://doi.org/10.1287/isre.6.2.144 Ajzen, “The theory of planned behavior,” Organizational Behavior and Human Decision Processes, vol. 50, no. 2, pp. 179–211, 1991. [Online]. Available: https://doi.org/10.1016/0749-5978(91)90020-T Cabero-Almenara, A. Palacios-Rodríguez, H. A. Rojas Guzmán, and V. Fernández-Scagliusi, “Prediction of the use of generative artificial intelligence through ChatGPT among Costa Rican university students: A PLS model based on UTAUT2,” Applied Sciences, vol. 15, no. 6, p. 3363, 2025. [Online]. Available: https://doi.org/10.3390/app15063363 M. Al-Okaily, W. Mater, N. Matar, and F. S. Shiyyab, “The role of ChatGPT adoption in enhancing communication and collaboration skills among business school students,” Computers and Education: Artificial Intelligence, p. 100441, 2025. [Online]. Available: https://doi.org/10.1016/j.caeai.2025.100441 Y. Zhang and W. Pan, "A UTAUT2 model expansion: Investigating the effect of interactive learning environment and gamification on Duolingo user base in China," International Journal of Human–Computer Interaction , vol. 41, no. 15, pp. 9518–9532, 2025. [Online]. Available: https://doi.org/10.1080/10447318.2024.2426736 Y. Zhang and Z. Yu, "Emotional attachment as the key mediator: Expanding UTAUT2 to examine how perceived anthropomorphism in intelligent agents influences the sustained use of DouBao (Cici) among EFL learners," Education and Information Technologies , vol. 30, pp. 23897–23919, 2025. [Online]. Available: https://doi.org/10.1007/s10639-025-13721-3 X. Deng and Z. Yu, "An extended hedonic motivation adoption model of TikTok in higher education," Education and Information Technologies , vol. 28, no. 10, pp. 13595–13617, 2023. [Online]. Available: https://doi.org/10.1007/s10639-023-11749-x Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 14 Apr, 2026 Reviewers agreed at journal 10 Mar, 2026 Reviews received at journal 06 Mar, 2026 Reviewers agreed at journal 06 Mar, 2026 Reviews received at journal 08 Feb, 2026 Reviewers agreed at journal 06 Feb, 2026 Reviewers invited by journal 06 Feb, 2026 Editor assigned by journal 16 Jan, 2026 Submission checks completed at journal 14 Jan, 2026 First submitted to journal 14 Jan, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8564643","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":589853002,"identity":"6ead8321-120d-46c8-a598-c0a3133047f8","order_by":0,"name":"Faisal Alshammari","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA60lEQVRIie3OIQvCQBTA8XccuHLDOovfQHhVRP0gFkHQsoFdQdOtqNmvsXIYbxxomVgXXVkyLIkr4qmYxEOb4f5l3LHfvQdgs/1hzhyIJPPnQRaPT99MmNR/vki8/plQ9hUBeozLTafXCPeZavsKqo6P5DI1kQoqNxkEIhmhCoSC2uKElG0/ky4AKsJpIOQQHgRTHylUTFOcIi75LBCHHFRTk256X+xqIgyly/X7qZ5C7lM8H8HlBkLZWLl8p0kO8VKMmJfk+mZlIE4YZSWf6MWGtChFq14NB1F2OX8mQN821UkDsNlsNtsX3QDvMVa0lyoP3wAAAABJRU5ErkJggg==","orcid":"","institution":"Majmaah University","correspondingAuthor":true,"prefix":"","firstName":"Faisal","middleName":"","lastName":"Alshammari","suffix":""}],"badges":[],"createdAt":"2026-01-09 23:38:12","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8564643/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8564643/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102743323,"identity":"4fd88ef8-6502-4fc8-b3d0-ae9ec147b480","added_by":"auto","created_at":"2026-02-16 08:21:17","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":99454,"visible":true,"origin":"","legend":"\u003cp\u003eConceptual landscape of ChatGPT adoption in programming education. Note: This figure presents an integrative overview for orientation purposes. The formal research model (Figure 2) specifies testable hypotheses. Gender moderates the PEOU→BI and ANX→BI paths; IT Experience moderates the PU→BI and AIL→BI paths.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8564643/v1/ce1f8b147d6d176f6d978fdd.png"},{"id":102743328,"identity":"55862722-d86a-4d04-9063-a9175b14e14d","added_by":"auto","created_at":"2026-02-16 08:21:17","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":515772,"visible":true,"origin":"","legend":"\u003cp\u003eAdoptGPT-Prog research model showing hypothesized relationships.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8564643/v1/a5016991802979825eee66d9.png"},{"id":102749396,"identity":"3c3d2949-f3e8-41f4-9e01-6b193c4eff90","added_by":"auto","created_at":"2026-02-16 09:12:34","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":473907,"visible":true,"origin":"","legend":"\u003cp\u003eAdoptGPT-Prog research framework showing direct effects (H1–H8) and moderation effects of Gender (H9a–H9c) and IT Experience (H10a–H10c) on Behavioral Intention\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8564643/v1/3244326c7c8376893b851ccc.png"},{"id":102743322,"identity":"d60531a6-df6e-4c40-8da2-6becb203fc5e","added_by":"auto","created_at":"2026-02-16 08:21:17","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":492880,"visible":true,"origin":"","legend":"\u003cp\u003eStructural Model Results with Standardized Path Coefficients\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8564643/v1/3c286112fbee547e85178d53.png"},{"id":102743324,"identity":"8bbbe04f-571c-456b-96a2-72e361b7485b","added_by":"auto","created_at":"2026-02-16 08:21:17","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":239474,"visible":true,"origin":"","legend":"\u003cp\u003eGender Moderation Analysis\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8564643/v1/4db2b5d2cb0284d7935430ea.png"},{"id":102743325,"identity":"d8d2287e-316a-43b7-80a7-f71a6a78670c","added_by":"auto","created_at":"2026-02-16 08:21:17","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":233406,"visible":true,"origin":"","legend":"\u003cp\u003eIT Experience Moderation Analysis\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8564643/v1/7d85d92680d64c1e0e4f08da.png"},{"id":102749341,"identity":"f84ccd75-9676-4b5c-b52c-1e63e272f898","added_by":"auto","created_at":"2026-02-16 09:12:24","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":101316,"visible":true,"origin":"","legend":"\u003cp\u003eImportance-Performance Map Analysis (IPMA)\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8564643/v1/79e9ffd162e0a1462bd6076e.png"},{"id":102749014,"identity":"c4bcd86e-d078-4168-a0ff-85d6f75c2209","added_by":"auto","created_at":"2026-02-16 09:11:51","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":245108,"visible":true,"origin":"","legend":"\u003cp\u003eEffect Size Comparison\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8564643/v1/e98c1a643b19bed9d84f1de9.png"},{"id":103049131,"identity":"07c4a908-afb0-4510-959c-727f04afc658","added_by":"auto","created_at":"2026-02-20 07:33:54","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4655354,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8564643/v1/d6def8a2-9d8f-42f8-a7db-9a5f80504ef7.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Understanding Factors Influencing Students Intention to Use ChatGPT for Learning Programming with Gender and IT Experience as Moderators Based on AdoptGPT Prog Conceptual Model","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eThe integration of generative artificial intelligence (AI) into higher education has created unprecedented opportunities and challenges, particularly in programming education where students must develop complex cognitive skills including debugging, algorithmic thinking, and code comprehension [1]. ChatGPT, developed by OpenAI, represents a significant advancement in natural language processing that enables conversational interaction and code generation, fundamentally altering how students can engage with programming concepts [2]. However, despite widespread interest in these AI-assisted learning tools, a critical question remains inadequately addressed: What factors determine whether programming students will adopt ChatGPT as a learning resource, and how do individual differences shape these adoption decisions?\u003c/p\u003e \u003cdiv id=\"Sec2\" class=\"Section2\"\u003e \u003ch2\u003e1.1 Problem Statement and Research Gap\u003c/h2\u003e \u003cp\u003eExisting research on ChatGPT adoption in higher education has predominantly applied generic technology acceptance frameworks without accounting for the unique cognitive demands of programming education [3\u0026ndash;8]. Programming differs fundamentally from other educational domains\u0026mdash;it requires iterative problem-solving, tolerance for ambiguity in debugging, and the ability to evaluate AI-generated code for correctness and efficiency [4, 5]. These domain-specific characteristics suggest that adoption factors may operate differently in programming contexts compared to general educational technology adoption.\u003c/p\u003e \u003cp\u003e \u003cem\u003eThree specific gaps motivate this research\u003c/em\u003e:\u003c/p\u003e \u003cp\u003eFirst, while TAM and UTAUT frameworks have been extensively applied to educational technology [7, 14], their application to ChatGPT in programming education remains limited. Recent studies by Al-Sharafi et al. [17] and Strzelecki [28] examined ChatGPT adoption in general higher education contexts, but neither specifically addressed programming education nor examined how the technical nature of programming tasks might differentially influence adoption factors. Similarly, Balaskas et al. examined ChatGPT adoption but focused on general academic use rather than discipline-specific applications.\u003c/p\u003e \u003cp\u003eSecond, demographic moderators\u0026mdash;particularly gender and IT experience\u0026mdash;have been examined separately in technology adoption research [10, 11, 22, 23], but their combined moderating effects on ChatGPT adoption in a technical domain remain unexplored. This oversight is consequential because programming education exhibits persistent gender disparities [22], and students enter programming courses with vastly different technical backgrounds [24], suggesting that adoption pathways may vary systematically across these demographic dimensions.\u003c/p\u003e \u003cp\u003eThird, existing ChatGPT adoption models have not adequately distinguished between cognitive-affective factors (trust, anxiety, perceived risk) and competency factors (AI literacy, perceived ease of use) that may differentially predict adoption intention in technically demanding contexts. Understanding this distinction is essential for developing targeted interventions that address the specific barriers faced by different student populations.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e1.2 Research Questions\u003c/h2\u003e \u003cp\u003eTo address these gaps, this study investigates the following research questions:\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eRQ1\u003c/strong\u003e \u003cp\u003eWhat is the relative importance of cognitive-affective factors (perceived usefulness, trust, anxiety, perceived risk) versus competency factors (AI literacy, perceived ease of use, perceived learning value) in predicting programming students' intention to adopt ChatGPT?\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eRQ2\u003c/strong\u003e \u003cp\u003eDoes gender moderate the relationships between adoption factors and behavioral intention, and if so, which specific pathways exhibit gender differences?\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eRQ3\u003c/strong\u003e \u003cp\u003eHow does prior IT experience moderate the relationship between perceived benefits (perceived usefulness, AI literacy) and adoption intention among programming students?\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eRQ4\u003c/strong\u003e \u003cp\u003eDoes the proposed integrated TAM-UTAUT framework (AdoptGPT-Prog) achieve superior explanatory power compared to existing ChatGPT adoption models when applied to programming education?\u003c/p\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e1.3 Research Objectives and Approach\u003c/h2\u003e \u003cp\u003eThis study proposes AdoptGPT-Prog, an integrated TAM-UTAUT framework specifically designed for understanding ChatGPT adoption in programming education. The model incorporates eight direct predictors organized into two theoretically meaningful categories:\u003c/p\u003e \u003cp\u003eEnablers: Perceived Usefulness (PU), Perceived Ease of Use (PEOU), Trust (TR), Perceived Learning Value (PLV), AI Literacy (AIL), and Hedonic Motivation (HM)\u003c/p\u003e \u003cp\u003eBarriers: Perceived Risk (PR) and Anxiety (ANX)\u003c/p\u003e \u003cp\u003eGender and IT Experience function as moderators on theoretically selected pathways, as detailed in Section 3.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presents the conceptual landscape that situates this research within the broader technology adoption literature, illustrating how technological perceptions, individual characteristics, and contextual elements interact in ChatGPT adoption decisions. This conceptual mapping orients readers to the multifaceted nature of the phenomenon under investigation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e1.4 Contributions\u003c/h2\u003e \u003cp\u003eThis research makes four distinct contributions:\u003c/p\u003e \u003cp\u003e \u003col\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eDomain-Specific Theoretical Extension\u003c/b\u003e: Unlike prior studies that applied generic TAM-UTAUT frameworks to ChatGPT adoption [17, 28], AdoptGPT-Prog explicitly accounts for the cognitive demands of programming education by distinguishing enabler and barrier constructs and examining their relative importance in a technical learning context.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eSystematic Moderator Analysis\u003c/b\u003e: This study provides the first empirical examination of gender and IT experience as simultaneous moderators of ChatGPT adoption in programming education, with theory-driven selection of moderated pathways (detailed in Section \u003cspan refid=\"Sec22\" class=\"InternalRef\"\u003e3.4\u003c/span\u003e).\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eValidated Domain-Specific Instrument\u003c/b\u003e: The study develops and validates a measurement instrument tailored to ChatGPT adoption in programming contexts, with items addressing programming-specific applications (debugging assistance, code generation, algorithm explanation).\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003cspan\u003e \u003cli\u003e \u003cp\u003e \u003cb\u003eDifferentiated Practical Recommendations\u003c/b\u003e: Based on moderation analyses, this study provides actionable recommendations stratified by student gender and experience level, enabling educators to implement targeted interventions.\u003c/p\u003e \u003c/li\u003e \u003c/span\u003e \u003c/ol\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e1.5 Paper Organization\u003c/h2\u003e \u003cp\u003eThe remainder of this paper is organized as follows. Section 2 presents an integrated literature review covering technology acceptance frameworks, ChatGPT adoption research, and demographic moderators. Section 3 describes the theoretical framework, hypothesis development, and methodology. Section 4 reports results including measurement validation, structural model analysis, and moderation effects. Section \u003cspan refid=\"Sec35\" class=\"InternalRef\"\u003e5\u003c/span\u003e discusses findings in relation to prior research. Section \u003cspan refid=\"Sec43\" class=\"InternalRef\"\u003e6\u003c/span\u003e addresses implications, limitations, and future directions. Section \u003cspan refid=\"Sec46\" class=\"InternalRef\"\u003e7\u003c/span\u003e concludes with key contributions and recommendations.\u003c/p\u003e \u003c/div\u003e"},{"header":"2. Theoretical Foundation and Literature Review","content":"\u003cp\u003eThis section provides an integrated review of technology acceptance frameworks, ChatGPT adoption research, and the theoretical foundations for construct selection and hypothesis development. Rather than presenting constructs in isolation, we organize the literature around three themes: (1) foundational acceptance theories, (2) ChatGPT adoption in educational contexts, and (3) the rationale for construct selection and moderator specification in programming education.\u003c/p\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n \u003ch2\u003e2.1 Technology Acceptance Frameworks: From TAM to Integrated Models\u003c/h2\u003e\n \u003cp\u003eThe Technology Acceptance Model (TAM), proposed by Davis [7], established perceived usefulness (PU) and perceived ease of use (PEOU) as fundamental determinants of technology adoption intention. Venkatesh et al. [13] extended this foundation through the Unified Theory of Acceptance and Use of Technology (UTAUT), incorporating performance expectancy, effort expectancy, social influence, and facilitating conditions as core constructs, with gender, age, experience, and voluntariness as moderators.\u003c/p\u003e\n \u003cp\u003eMeta-analytic evidence confirms the robustness of these frameworks in educational technology contexts, with explained variance in behavioral intention typically ranging from 40\u0026ndash;70% [14, 15]. However, recent scholarship advocates for integrated TAM-UTAUT approaches when examining emerging technologies, as single-framework applications may inadequately capture the complexity of novel AI-assisted learning tools [16]. This integration strategy informs our theoretical approach.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\n \u003ch2\u003e2.2 ChatGPT Adoption Research: Current Evidence and Limitations\u003c/h2\u003e\n \u003cp\u003eSince ChatGPT\u0026apos;s public release in November 2022, adoption research has proliferated across educational contexts. Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e summarizes key studies and positions our contribution relative to existing work.\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eComparison of ChatGPT Adoption Studies in Higher Education\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"8\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStudy\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eFramework\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eContext\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSample\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eKey Predictors (\u0026beta;)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eR\u0026sup2;\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eModerators\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eLimitations\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBalaskas et al. [8]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTAM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGreek universities (general)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e412\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePU (0.42), Attitude (0.38)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.58\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNo domain specificity; no demographic moderators\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAl-Sharafi et al. [17]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMeta-UTAUT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDutch higher education\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e387\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePE (0.35), Attitude (0.31), SI (0.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGeneral academic use; Western context only\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eStrzelecki [28]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eExtended UTAUT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePolish universities\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e534\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePE (0.38), HM (0.24), SI (0.19)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNone\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNo programming focus; limited construct set\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eYilmaz et al. [9]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDOI Theory\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTurkish universities\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e298\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRelative advantage, compatibility\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGender (partial)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDifferent theoretical base; no IT experience\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eThis study\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eIntegrated TAM-UTAUT\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSaudi Arabia (programming)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e486\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePU, PEOU, TR, PLV, AIL, PR, ANX, HM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.714\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGender, IT Experience\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eDomain-specific; dual moderators\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003e\u003cem\u003eThree limitations emerge from existing research that motivate our study\u003c/em\u003e:\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eLimitation 1\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eLack of Domain Specificity. Prior studies examine ChatGPT adoption in general academic contexts without accounting for discipline-specific factors. Programming education involves unique cognitive demands\u0026mdash;debugging, algorithm design, code evaluation\u0026mdash;that may differentially weight adoption factors. For instance, trust in AI-generated code correctness may be more consequential than trust in AI-generated essay content.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eLimitation 2\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eIncomplete Moderator Examination. While Yilmaz et al. [9] found partial gender effects (males emphasized compatibility; females emphasized trialability), no study has simultaneously examined gender and IT experience as moderators within an integrated framework. Given persistent gender disparities in computing education [22] and the technical heterogeneity among programming students [25], this oversight limits practical applicability.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eLimitation 3\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eNarrow Construct Sets. Existing models typically include 4\u0026ndash;6 predictors, potentially omitting factors relevant to AI-assisted programming learning. Constructs such as AI literacy, perceived learning value, and perceived risk have theoretical relevance but remain underexplored in ChatGPT adoption contexts.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\n \u003ch2\u003e2.3 Programming Education Context: Why Domain Specificity Matters\u003c/h2\u003e\n \u003cp\u003eProgramming education presents distinctive characteristics that justify domain-specific model development:\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eCognitive Complexity\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eProgramming requires iterative problem-solving, tolerance for ambiguity during debugging, and the ability to evaluate code correctness\u0026mdash;skills that interact with AI tool adoption in ways not captured by general educational technology models [4, 19].\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eOutput Verifiability\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eUnlike essay writing or information retrieval, AI-generated code can be objectively tested for correctness through compilation and execution. This verifiability may heighten the salience of trust and perceived risk compared to other educational applications.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eTechnical Prerequisites\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eProgramming students possess varying levels of prior technical experience, creating heterogeneity in how they evaluate and interact with AI tools [21, 25]. This variability supports IT experience as a theoretically meaningful moderator.\u003c/p\u003e\n \u003cp\u003eEmpirical evidence from programming-specific studies supports these distinctions. Sun et al. [4] found that ChatGPT-facilitated programming (CFP) learners exhibited different behavioral patterns than self-directed programmers, with CFP students showing enhanced debugging behaviors and improved usefulness perceptions. Xiao et al. [20] demonstrated that structured prompting improved interaction quality in programming tasks, suggesting that AI literacy\u0026mdash;the ability to craft effective prompts\u0026mdash;may be particularly consequential in this domain. Silva et al. [21] identified prior knowledge and cognitive comprehension as moderating the effectiveness of ChatGPT for programming problem-solving.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n \u003ch2\u003e2.4 Construct Selection and Theoretical Rationale\u003c/h2\u003e\n \u003cp\u003eBased on the foregoing review, we organize the AdoptGPT-Prog model around two construct categories\u0026mdash;enablers and barriers\u0026mdash;reflecting their hypothesized directional effects on adoption intention.\u003c/p\u003e\n \u003cdiv id=\"Sec12\" class=\"Section3\"\u003e\n \u003ch2\u003e2.4.1 Enablers: Factors Facilitating Adoption\u003c/h2\u003e\n \u003cp\u003ePerceived Usefulness (PU) represents the degree to which students believe ChatGPT will enhance their programming performance. Drawing on TAM [7] and UTAUT\u0026apos;s performance expectancy construct [13], PU consistently emerges as the strongest adoption predictor across educational technologies. In programming contexts, usefulness perceptions encompass debugging assistance, code generation efficiency, and concept clarification capabilities.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH1\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003ePerceived Usefulness positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e\n \u003cp\u003ePerceived Ease of Use (PEOU) captures students\u0026apos; assessments of the cognitive effort required to interact with ChatGPT effectively. TAM posits PEOU as both a direct predictor of intention and an indirect predictor through PU [7]. For programming students, ease of use encompasses prompt formulation, output interpretation, and integration into existing workflows.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH2\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003ePerceived Ease of Use positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e\n \u003cp\u003eTrust (TR) refers to students\u0026apos; confidence that ChatGPT provides accurate, reliable, and unbiased responses. Trust is particularly consequential in AI-assisted learning contexts where students rely on AI outputs for understanding and decision-making [17, 33]. In programming education, trust encompasses confidence in code correctness, explanation accuracy, and debugging recommendations. Students who trust ChatGPT\u0026apos;s outputs are more likely to integrate it into their learning practices.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH3\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eTrust positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e\n \u003cp\u003ePerceived Learning Value (PLV) represents students\u0026apos; beliefs that ChatGPT meaningfully enhances their programming knowledge, skills, and academic outcomes. Al-Okaily et al. [33] demonstrate that perceived educational benefits drive adoption intention for generative AI tools. In programming courses, learning value perceptions arise from syntax clarification, debugging assistance, and algorithmic logic explanations that complement traditional instruction.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH4\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003ePerceived Learning Value positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e\n \u003cp\u003eAI Literacy (AIL) encompasses students\u0026apos; understanding of AI system capabilities, limitations, ethical considerations, and effective usage strategies. Higher AI literacy enables students to craft effective prompts, critically evaluate AI-generated code, and recognize potential errors [21, 33]. In programming contexts, AI-literate students can more effectively leverage ChatGPT\u0026apos;s capabilities while avoiding pitfalls such as uncritical code acceptance.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH5\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eAI Literacy positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e\n \u003cp\u003eHedonic Motivation (HM) captures the enjoyment, curiosity, and intrinsic interest students experience when interacting with ChatGPT. UTAUT2 established hedonic motivation as a significant adoption driver, particularly for consumer technologies [13]. ChatGPT\u0026apos;s conversational interface, immediate responsiveness, and novelty may generate positive affective responses that reinforce adoption intention. However, in utilitarian educational contexts, hedonic factors may play a secondary role compared to performance-oriented constructs.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH6\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eHedonic Motivation positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec13\" class=\"Section3\"\u003e\n \u003ch2\u003e2.4.2 Barriers: Factors Inhibiting Adoption\u003c/h2\u003e\n \u003cp\u003ePerceived Risk (PR) encompasses students\u0026apos; concerns about misinformation, privacy violations, academic integrity implications, and AI-generated errors. Prior research establishes perceived risk as negatively influencing both trust and behavioral intention [33]. In programming education, risk perceptions may arise from concerns about incorrect code recommendations, plagiarism detection, or over-reliance on AI assistance that impedes genuine skill development.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH7\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003ePerceived Risk negatively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e\n \u003cp\u003eAnxiety (ANX) refers to apprehension, discomfort, or fear experienced when interacting with AI systems. Technology anxiety has been established as an adoption barrier across educational technologies [33]. In programming contexts, anxiety may stem from uncertainty about AI-generated code correctness, fear of academic misconduct accusations, or concerns about skill atrophy through AI dependence.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH8\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eAnxiety negatively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\n \u003ch2\u003e2.5 Moderator Specification: Theory-Driven Path Selection\u003c/h2\u003e\n \u003cp\u003eRather than testing all possible moderating effects\u0026mdash;which would risk Type I error inflation and reduce statistical power we specify moderating hypotheses based on theoretical precedent and domain-specific rationale. Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e summarizes the selection logic.\u003c/p\u003e\n \u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eTheoretical Rationale for Moderator Path Selection\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"5\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eModerator\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSelected Path\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eRationale\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eExcluded Paths\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eExclusion Rationale\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGender\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePEOU \u0026rarr; BI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eVenkatesh et al. [13] found effort expectancy effects stronger for females; programming\u0026apos;s technical interface may amplify gender differences in usability sensitivity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePU \u0026rarr; BI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGender gap in performance expectancy has diminished over time [22]\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGender\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eANX \u0026rarr; BI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eConsistent evidence of gender differences in technology anxiety [10, 22]; female students may experience heightened AI-related apprehension in male-dominated programming contexts\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTR \u0026rarr; BI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLimited/inconsistent evidence for gender-based trust differences in technology adoption\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eIT Experience\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePU \u0026rarr; BI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUTAUT establishes experience as moderating performance expectancy effects; experienced users better evaluate performance benefits [13, 24]\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePEOU \u0026rarr; BI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eEase of use becomes less salient as experience increases; ceiling effects expected\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eIT Experience\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAIL \u0026rarr; BI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAI literacy translates more effectively into adoption intention when combined with technical competence to implement AI-informed strategies\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eANX \u0026rarr; BI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRange restriction expected in programming student sample; anxiety variance limited\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003e\u003cstrong\u003eGender Moderation Hypotheses:\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eDrawing on gender differences in technology adoption [10, 13, 22] and the specific affordances of programming education, we hypothesize:\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH9a\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eGender moderates the relationship between Perceived Ease of Use and Behavioral Intention, such that the effect is stronger for female students than male students.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eRationale\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eVenkatesh et al. [13] demonstrated that effort expectancy effects were consistently stronger for females across multiple technologies. In programming education, where students must navigate ChatGPT\u0026apos;s interface to formulate prompts and interpret code outputs, female students may place greater emphasis on usability when forming adoption intentions. Additionally, women in computing fields may be more sensitive to interaction quality given documented experiences of stereotype threat and belonging uncertainty [22].\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH9b\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eGender moderates the relationship between Anxiety and Behavioral Intention, such that the negative effect is stronger for female students than male students.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eRationale\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eResearch consistently documents gender differences in technology-related anxiety, with females reporting higher anxiety levels across educational technologies [10, 22]. In programming contexts\u0026mdash;a domain with persistent gender underrepresentation\u0026mdash;female students may experience amplified AI-related apprehension that more strongly inhibits adoption intention.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eIT Experience Moderation Hypotheses\u003c/strong\u003e:\u003c/p\u003e\n \u003cp\u003eBased on UTAUT\u0026apos;s experience moderator specification [13] and programming education\u0026apos;s technical demands [24, 25]:\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH10a\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eIT Experience moderates the relationship between Perceived Usefulness and Behavioral Intention, such that the effect is stronger for students with high IT experience.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eRationale\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eExperienced users possess cognitive frameworks that enable accurate assessment of how new tools enhance performance [13, 24]. In programming contexts, students with substantial IT backgrounds can better evaluate ChatGPT\u0026apos;s debugging, code generation, and explanation capabilities, translating usefulness perceptions more directly into adoption intention.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eH10b\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eIT Experience moderates the relationship between AI Literacy and Behavioral Intention, such that the effect is stronger for students with high IT experience.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eRationale\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eAI literacy\u0026mdash;understanding AI capabilities, limitations, and effective prompting\u0026mdash;becomes actionable when combined with technical competence. Students with high IT experience possess the programming skills necessary to implement AI-informed strategies (e.g., integrating AI-generated code snippets, debugging AI suggestions), amplifying the adoption benefits of AI knowledge.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\n \u003ch2\u003e2.6 Research Model Summary\u003c/h2\u003e\n \u003cp\u003eFigure \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e presents the complete AdoptGPT-Prog research model integrating the eight direct effect hypotheses (H1-H8) and four moderation hypotheses (H9a, H9b, H10a, H10b). The model organizes constructs into enablers (PU, PEOU, TR, PLV, AIL, HM) and barriers (PR, ANX), with Gender and IT Experience moderating theoretically specified pathways.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"3. Proposed Methodology","content":"\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003e3.1. Research Model Overview\u003c/h2\u003e \u003cp\u003eThe AdoptGPT-Prog model integrates constructs from the Technology Acceptance Model (TAM) and the Unified Theory of Acceptance and Use of Technology (UTAUT) to examine factors influencing students' intention to use ChatGPT for programming learning. Building on Davis's [7] foundational TAM framework and Venkatesh et al.'s [13] unified theory, this study develops a domain-specific model tailored to the unique demands of programming education.\u003c/p\u003e \u003cp\u003eThe model incorporates eight direct predictors of Behavioral Intention (BI), organized into two categories based on the extended AI adoption framework proposed by Al-Okaily et al. [33]:\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eEnablers\u003c/strong\u003e \u003cp\u003ePerceived Usefulness (PU), Perceived Ease of Use (PEOU), Trust (TR), Perceived Learning Value (PLV), AI Literacy (AIL), and Hedonic Motivation (HM)\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eBarriers\u003c/strong\u003e \u003cp\u003ePerceived Risk (PR) and Anxiety (ANX)\u003c/p\u003e \u003c/p\u003e \u003cp\u003eTwo moderating variables\u0026mdash;Gender and IT Experience\u0026mdash;are incorporated to examine demographic differences in adoption patterns. This approach aligns with Venkatesh et al. [13], who demonstrated that gender and experience significantly moderate technology acceptance relationships, and with subsequent meta-analyses confirming these effects in educational contexts [10, 14]. Rather than testing all possible moderation effects (which would yield 16 interaction terms and inflate Type I error rates), we adopted a theory-driven approach to select six moderation hypotheses based on theoretical foundations and empirical precedent.\u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e presents the complete AdoptGPT-Prog research framework illustrating all hypothesized direct effects (H1\u0026ndash;H8) and moderation effects (H9a\u0026ndash;H9c for gender; H10a\u0026ndash;H10c for IT experience).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Hypotheses Development\u003c/h2\u003e \u003cp\u003eBased on the integrated TAM-UTAUT framework and prior research on AI adoption in education [8, 17, 28, 33], this study proposes fourteen hypotheses organized into direct effects and moderation effects.\u003c/p\u003e \u003cdiv id=\"Sec19\" class=\"Section3\"\u003e \u003ch2\u003e3.2.1 Direct Effects on Behavioral Intention\u003c/h2\u003e \u003cp\u003eDrawing on TAM [7] and UTAUT [13], effort expectancy (perceived ease of use) directly influences users' intention to adopt new technologies. In programming education, where students must navigate complex syntax and debugging processes, the perceived simplicity of ChatGPT interactions becomes particularly salient [4, 19].\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH1\u003c/strong\u003e \u003cp\u003ePerceived Ease of Use positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e \u003c/p\u003e \u003cp\u003ePerceived risk captures concerns about misinformation, academic integrity, and privacy that may inhibit adoption [6, 8]. Balaskas et al. [8] found that risk perceptions significantly mediate the relationship between trust and adoption intention in higher education contexts.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH2\u003c/strong\u003e \u003cp\u003ePerceived Risk negatively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e \u003c/p\u003e \u003cp\u003eAI literacy\u0026mdash;understanding of AI capabilities, limitations, and responsible use\u0026mdash;enhances students' confidence in evaluating ChatGPT outputs [25, 33]. Al-Okaily et al. [33] demonstrated that AI literacy directly predicts adoption intention by reducing uncertainty and enabling effective tool utilization.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH3\u003c/strong\u003e \u003cp\u003eAI Literacy positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e \u003c/p\u003e \u003cp\u003eTrust represents confidence in ChatGPT's accuracy, reliability, and security [6, 17, 18]. Al-Sharafi et al. [17] found trust to be a cornerstone of ChatGPT adoption, while Niu and Mvondo [18] confirmed its central role in higher education acceptance.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH4\u003c/strong\u003e \u003cp\u003eTrust positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e \u003c/p\u003e \u003cp\u003eTechnology anxiety\u0026mdash;discomfort or apprehension when interacting with AI systems\u0026mdash;has been identified as a significant barrier to educational technology adoption [14, 26]. In the context of generative AI, anxiety may stem from uncertainty about output quality or fear of becoming overly dependent on AI assistance.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH5\u003c/strong\u003e \u003cp\u003eAnxiety negatively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e \u003c/p\u003e \u003cp\u003ePerceived usefulness (performance expectancy) is consistently the strongest predictor of technology adoption [7, 13, 14]. In programming education, ChatGPT's utility for code generation, debugging, and concept explanation directly addresses students' performance goals [4, 19, 20].\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH6\u003c/strong\u003e \u003cp\u003ePerceived Usefulness positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e \u003c/p\u003e \u003cp\u003ePerceived learning value captures beliefs about educational benefits\u0026mdash;enhanced understanding, skill development, and academic performance [33]. Zhang and Pan [34] demonstrated that learning value perceptions significantly predict adoption in educational technology contexts.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH7\u003c/strong\u003e \u003cp\u003ePerceived Learning Value positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e \u003c/p\u003e \u003cp\u003eHedonic motivation\u0026mdash;enjoyment, curiosity, and novelty appreciation\u0026mdash;influences adoption when technology use involves intrinsic satisfaction [32, 36]. Deng and Yu [36] found hedonic motivation significantly predicted educational technology adoption, while Cabero-Almenara et al. [32] confirmed its relevance for ChatGPT use among university students.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH8\u003c/strong\u003e \u003cp\u003eHedonic Motivation positively influences Behavioral Intention to use ChatGPT for programming learning.\u003c/p\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section3\"\u003e \u003ch2\u003e3.2.2 Moderation Effects\u003c/h2\u003e \u003cp\u003e \u003cb\u003eGender Moderation\u003c/b\u003e: The original UTAUT framework [13] established that gender moderates key technology acceptance relationships. Subsequent research has confirmed these effects: Cai et al. [10] meta-analyzed gender differences in technology attitudes, while Campos and Scherer [22] documented persistent digital gender gaps across 32 countries. Drawing on this evidence, three gender moderation hypotheses were specified for paths where theory and empirical precedent suggest differential effects.\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH9a\u003c/strong\u003e \u003cp\u003eGender moderates the relationship between Perceived Ease of Use and Behavioral Intention, such that the effect is stronger for female students.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eRationale\u003c/strong\u003e \u003cp\u003eVenkatesh et al. [13] demonstrated effort expectancy effects are stronger for women. This finding has been replicated in educational technology contexts [10, 11, 22], where female students place greater emphasis on usability factors.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH9b\u003c/strong\u003e \u003cp\u003eGender moderates the relationship between Perceived Risk and Behavioral Intention.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eRationale\u003c/strong\u003e \u003cp\u003eWomen exhibit higher risk sensitivity in technology-related decisions, particularly concerning privacy and data security [10, 11]. In AI-assisted programming, concerns about misinformation and academic integrity may manifest differently across genders.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH9c\u003c/strong\u003e \u003cp\u003eGender moderates the relationship between Anxiety and Behavioral Intention.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eRationale\u003c/strong\u003e \u003cp\u003eTechnology anxiety shows gender-differentiated effects in educational settings [10, 14]. Female students often report higher levels of computer anxiety, which may more strongly inhibit adoption intentions [22].\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eIT Experience Moderation\u003c/strong\u003e \u003cp\u003eUTAUT [13] established that experience moderates technology acceptance relationships. Experienced users possess better-developed mental models for evaluating new technologies [23, 24] and can more effectively leverage their technical knowledge [21, 25]. Three IT experience moderation hypotheses were specified based on theoretical predictions about how experience shapes evaluation processes.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH10a\u003c/strong\u003e \u003cp\u003eIT Experience moderates the relationship between Perceived Usefulness and Behavioral Intention, such that the effect is stronger for students with higher IT experience.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eRationale\u003c/strong\u003e \u003cp\u003eExperienced users can better assess productivity gains and recognize how ChatGPT capabilities translate to performance improvements [13, 24].\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH10b\u003c/strong\u003e \u003cp\u003eIT Experience moderates the relationship between AI Literacy and Behavioral Intention.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eRationale\u003c/strong\u003e \u003cp\u003eTechnical experience provides cognitive frameworks that amplify AI literacy benefits, enabling more effective translation of AI knowledge into usage strategies [21, 25].\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH10c\u003c/strong\u003e \u003cp\u003eIT Experience moderates the relationship between Trust and Behavioral Intention.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eRationale\u003c/strong\u003e \u003cp\u003eExperienced users may rely less on trust perceptions as they possess alternative evaluation mechanisms (direct assessment, technical understanding) [17, 18].\u003c/p\u003e \u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Rationale for Moderation Path Selection\u003c/h2\u003e \u003cp\u003eThe selective approach to moderation testing follows best practices in structural equation modeling, which recommend theory-driven hypothesis specification over exhaustive testing to maintain parsimony and control Type I error [14]. Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e summarizes the rationale for path selection and exclusion decisions.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eModeration Path Selection Rationale\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eModerator\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePath\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eRationale for Selection/Exclusion\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePEOU\u0026rarr;BI ✓\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSelected: UTAUT [13] shows effort expectancy effects stronger for women; replicated in educational contexts [10, 22]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePR\u0026rarr;BI ✓\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSelected: Women exhibit higher risk sensitivity in technology decisions [10, 11]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eANX\u0026rarr;BI ✓\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSelected: Gender-differentiated anxiety effects documented in educational settings [10, 14]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTR\u0026rarr;BI ✗\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eExcluded: Limited evidence for gender differences in trust formation toward AI [8, 9, 17]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePU\u0026rarr;BI ✗\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eExcluded: Gender difference in performance expectancy has diminished [14]; recent ChatGPT studies show no significant difference [8, 9]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIT Experience\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePU\u0026rarr;BI ✓\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSelected: Experienced users better assess productivity gains [13, 24, 25]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIT Experience\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAIL\u0026rarr;BI ✓\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSelected: Technical experience amplifies AI literacy benefits [21, 25]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIT Experience\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTR\u0026rarr;BI ✓\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSelected: Experienced users have alternative evaluation mechanisms beyond trust [17, 18]\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIT Experience\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePEOU\u0026rarr;BI ✗\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eExcluded: UTAUT [13] suggests ease of use becomes less salient with experience*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"3\"\u003eNote: ✓ = Selected for hypothesis testing; ✗ = Excluded based on theoretical/empirical rationale. *This path emerged as significant in exploratory multi-group analysis and is reported in Results.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003e3.4 Data Collection and Sample\u003c/h2\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003e3.4.1 Participants and Procedure\u003c/h2\u003e \u003cp\u003eData were collected through an online survey administered to undergraduate programming students at five universities in Saudi Arabia during March-April 2024. Stratified random sampling ensured balanced representation across gender and IT experience levels. The survey was available in both English and Arabic to accommodate participant preferences.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section3\"\u003e \u003ch2\u003e3.4.2 Inclusion and Exclusion Criteria\u003c/h2\u003e \u003cp\u003e \u003cb\u003eInclusion Criteria\u003c/b\u003e:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eCurrently enrolled undergraduate in computer science, IT, software engineering, or related disciplines\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eActive enrollment in programming course (Python, Java, C++, or JavaScript)\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003ePrior awareness and at least minimal exposure to ChatGPT\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eAged 18 years or older with voluntary informed consent\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eExclusion Criteria\u003c/b\u003e:\u003c/p\u003e \u003cp\u003e \u003cul\u003e \u003cli\u003e \u003cp\u003eIncomplete responses (\u0026gt;\u0026thinsp;10% missing items) or straight-lining patterns\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eFailed attention check questions\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eStatistical outliers identified through Mahalanobis distance (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001)\u003c/p\u003e \u003c/li\u003e \u003cli\u003e \u003cp\u003eDuplicate submissions identified via IP address verification\u003c/p\u003e \u003c/li\u003e \u003c/ul\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec25\" class=\"Section3\"\u003e \u003ch2\u003e3.4.3 Sample Characteristics\u003c/h2\u003e \u003cp\u003eFrom 523 initial responses, systematic screening yielded 486 valid responses (93% retention rate). The final sample comprised 52.3% male (n\u0026thinsp;=\u0026thinsp;254) and 47.7% female (n\u0026thinsp;=\u0026thinsp;232) participants, with a mean age of 21.4 years (SD\u0026thinsp;=\u0026thinsp;2.1, range 18\u0026ndash;28). Regarding IT experience, 38.5% (n\u0026thinsp;=\u0026thinsp;187) reported high experience, while 61.5% (n\u0026thinsp;=\u0026thinsp;299) reported low-to-moderate experience. Participants were distributed across five Saudi Arabian universities with balanced institutional representation.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec26\" class=\"Section2\"\u003e \u003ch2\u003e3.5 Measurement Instrument\u003c/h2\u003e \u003cp\u003eThe survey instrument contained 42 items measuring eight constructs on 7-point Likert scales (1\u0026thinsp;=\u0026thinsp;Strongly Disagree to 7\u0026thinsp;=\u0026thinsp;Strongly Agree). Items were adapted from validated scales: Perceived Usefulness and Perceived Ease of Use from Davis [7]; Performance Expectancy and Effort Expectancy from Venkatesh et al. [13]; Social Influence and Facilitating Conditions from Thompson et al. [29]; Attitude from Taylor and Todd [30]; Behavioral Intention from Ajzen [31]; and Trust, Anxiety, Perceived Learning Value, AI Literacy, Perceived Risk, and Hedonic Motivation from Al-Okaily et al. [33]. All items were adapted to the ChatGPT programming education context.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section2\"\u003e \u003ch2\u003e3.6 Data Analysis Approach\u003c/h2\u003e \u003cp\u003e Partial Least Squares Structural Equation Modeling (PLS-SEM) was employed using SmartPLS 4.0, following established guidelines [14]. PLS-SEM was selected over covariance-based SEM due to its suitability for exploratory research with complex models and its robustness with non-normal data distributions. Analysis proceeded in two stages:\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eStage 1 \u0026ndash; Measurement Model Assessment\u003c/strong\u003e \u003cp\u003eInternal consistency was evaluated through Cronbach's alpha and composite reliability (CR), with thresholds\u0026thinsp;\u0026ge;\u0026thinsp;0.70. Convergent validity was assessed via Average Variance Extracted (AVE\u0026thinsp;\u0026ge;\u0026thinsp;0.50) and indicator loadings (\u0026ge;\u0026thinsp;0.70). Discriminant validity was established using the Heterotrait-Monotrait (HTMT) ratio (\u0026lt;\u0026thinsp;0.85) and the Fornell-Larcker criterion.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eStage 2 \u0026ndash; Structural Model Assessment\u003c/strong\u003e \u003cp\u003ePath coefficients were tested for significance using bootstrapping with 5,000 resamples. Model explanatory power was assessed through R\u0026sup2; values (substantial\u0026thinsp;\u0026ge;\u0026thinsp;0.26, moderate\u0026thinsp;\u0026ge;\u0026thinsp;0.13, weak\u0026thinsp;\u0026ge;\u0026thinsp;0.02). Effect sizes (f\u0026sup2;) were calculated for each predictor (large\u0026thinsp;\u0026ge;\u0026thinsp;0.35, medium\u0026thinsp;\u0026ge;\u0026thinsp;0.15, small\u0026thinsp;\u0026ge;\u0026thinsp;0.02). Predictive relevance was established through Stone-Geisser's Q\u0026sup2; (\u0026gt;\u0026thinsp;0). Model fit was evaluated using SRMR (\u0026lt;\u0026thinsp;0.08).\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eModeration Analysis\u003c/strong\u003e \u003cp\u003eMulti-group analysis (MGA) was conducted to test gender and IT experience moderation effects. Groups were compared using permutation testing (5,000 permutations) and parametric difference tests. Significant moderation was established when path coefficient differences between groups exceeded critical thresholds (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). This approach follows recommendations by Venkatesh et al. [13] for examining demographic moderators in technology acceptance research.\u003c/p\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"4. Results and Evaluation","content":"\u003cdiv id=\"Sec29\" class=\"Section2\"\u003e \u003ch2\u003e4.1 Measurement Model Assessment\u003c/h2\u003e \u003cp\u003e Prior to hypothesis testing, the measurement model was evaluated for reliability and validity following established PLS-SEM guidelines [14]. Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e presents the construct reliability and convergent validity results.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eConstruct Reliability and Convergent Validity\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConstruct\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eItems\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCronbach's α\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eCR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAVE\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerceived Usefulness (PU)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.912\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.938\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.791\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerceived Ease of Use (PEOU)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.878\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.916\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.732\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTrust (TR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.889\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.923\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.750\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerceived Learning Value (PLV)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.875\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.914\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.727\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAI Literacy (AIL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.863\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.907\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.709\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHedonic Motivation (HM)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.858\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.903\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.701\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerceived Risk (PR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.851\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.899\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.691\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnxiety (ANX)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.847\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.897\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.686\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBehavioral Intention (BI)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.921\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.950\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.863\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003e\u003cem\u003eNote: CR\u0026thinsp;=\u0026thinsp;Composite Reliability; AVE\u0026thinsp;=\u0026thinsp;Average Variance Extracted. Thresholds: α\u0026thinsp;\u0026gt;\u0026thinsp;0.70, CR\u0026thinsp;\u0026gt;\u0026thinsp;0.70, AVE\u0026thinsp;\u0026gt;\u0026thinsp;0.50 [14].\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e presents the reliability and convergent validity assessment for all nine constructs in the AdoptGPT-Prog model. Internal consistency was evaluated using Cronbach's alpha (α) and Composite Reliability (CR), while convergent validity was assessed through Average Variance Extracted (AVE). All constructs demonstrated excellent internal consistency with Cronbach's α values ranging from 0.847 (Anxiety) to 0.921 (Behavioral Intention), exceeding the recommended threshold of 0.70. Composite Reliability values ranged from 0.897 to 0.950, further confirming strong internal consistency. Convergent validity was established with AVE values ranging from 0.686 (Anxiety) to 0.863 (Behavioral Intention), all exceeding the 0.50 threshold, indicating that each construct captures more than half of the variance in its indicators. These results confirm that the measurement model is psychometrically sound and suitable for structural model analysis.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eHeterotrait-Monotrait (HTMT) Ratio Matrix\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePU\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePEOU\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eTR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePLV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eAIL\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHM\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\"\u003e \u003cp\u003ePR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c9\"\u003e \u003cp\u003eANX\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePEOU\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.645\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eTR\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.612\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.489\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePLV\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.723\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.578\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.545\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eAIL\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.589\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.534\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.567\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.634\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eHM\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.534\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.512\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.423\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.612\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.456\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePR\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.334\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.312\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.398\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.356\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.287\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.278\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eANX\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.367\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.345\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.412\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.389\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.298\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.334\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.623\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e\u0026ndash;\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBI\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.756\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.623\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.656\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.689\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.578\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.534\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.445\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.489\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003e\u003cem\u003eNote: All HTMT values\u0026thinsp;\u0026lt;\u0026thinsp;0.85 threshold, confirming discriminant validity.\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e displays the Heterotrait-Monotrait (HTMT) ratio matrix used to assess discriminant validity among the nine constructs. HTMT values represent the ratio of between-trait correlations to within-trait correlations, with values below 0.85 indicating adequate discriminant validity. All construct pairs demonstrated HTMT values below this threshold, confirming that each construct is empirically distinct from others. The highest correlations were observed between Perceived Usefulness and Behavioral Intention (0.756) and between Perceived Usefulness and Perceived Learning Value (0.723), reflecting their conceptual relatedness while remaining within acceptable discriminant validity bounds. Notably, the correlation between Perceived Risk and Anxiety (0.623) confirms these barrier constructs are related but represent distinct phenomena\u0026mdash;risk captures cognitive concerns about potential negative outcomes, while anxiety reflects affective discomfort with AI interaction. These results support the theoretical distinction between all constructs in the model.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec30\" class=\"Section2\"\u003e \u003ch2\u003e4.2 Structural Model and Hypothesis Testing\u003c/h2\u003e \u003cp\u003eThe structural model explained 71.4% of variance in Behavioral Intention (R\u0026sup2; = 0.714), indicating substantial explanatory power that exceeds comparable studies [8, 17, 28]. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e presents the structural model with standardized path coefficients.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e illustrates the structural model results of AdoptGPT-Prog with standardized path coefficients for all hypothesized relationships. The figure displays eight predictor constructs organized into enablers (Perceived Usefulness, Perceived Ease of Use, Trust, Perceived Learning Value, AI Literacy, Hedonic Motivation) shown in blue boxes and barriers (Perceived Risk, Anxiety) shown in red boxes, all connected to the central dependent variable Behavioral Intention. Path coefficients are displayed on each arrow with significance indicators (***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001, **p\u0026thinsp;\u0026lt;\u0026thinsp;0.01, *p\u0026thinsp;\u0026lt;\u0026thinsp;0.05, ns\u0026thinsp;=\u0026thinsp;not significant). The strongest path emerges from Perceived Usefulness to Behavioral Intention (β\u0026thinsp;=\u0026thinsp;0.312***), highlighted with a green arrow, followed by Trust (β\u0026thinsp;=\u0026thinsp;0.203***), Perceived Learning Value and Perceived Ease of Use (both β\u0026thinsp;=\u0026thinsp;0.187***). Perceived Risk shows a significant negative effect (β = -0.156**), while Anxiety's path is marked as non-significant (β\u0026thinsp;=\u0026thinsp;0.012, ns). Moderation effects are depicted in yellow boxes showing Gender moderating PEOU\u0026rarr;BI (supported), PR\u0026rarr;BI (not supported), and ANX\u0026rarr;BI (supported), and IT Experience moderating PU\u0026rarr;BI (supported), AIL\u0026rarr;BI (supported), and TR\u0026rarr;BI (not supported). The model explains 71.4% of variance in Behavioral Intention (R\u0026sup2; = 0.714), as indicated in the BI construct box.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab6\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eHypothesis Testing Results\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePath\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003et-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eResult\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePEOU \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.187\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.845\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePR \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.156\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.234\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAIL \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.134\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.756\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.006\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTR \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.203\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e4.123\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eANX \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.289\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.774\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNot Supported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePU \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.312\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e6.478\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePLV \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.187\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.567\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHM \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.098\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.012\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.045\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH9a\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGender \u0026times; PEOU \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.167\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.345\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH9b\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGender \u0026times; PR \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.089\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.756\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.079\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNot Supported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH9c\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eGender \u0026times; ANX \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.112\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.234\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.026\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH10a\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eITE \u0026times; PU \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.178\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e3.567\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH10b\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eITE \u0026times; AIL \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.145\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e2.891\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eSupported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eH10c\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eITE \u0026times; TR \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.078\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.534\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.125\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNot Supported\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e\u003cem\u003eNote: β\u0026thinsp;=\u0026thinsp;standardized path coefficient. Results based on 5,000 bootstrap samples.\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e summarizes the hypothesis testing results for all fourteen hypotheses, including eight direct effects (H1\u0026ndash;H8) and six moderation effects (H9a\u0026ndash;H10c). Path coefficients (β), t-values, p-values, and support status are reported based on 5,000 bootstrap samples. Among direct effects, seven of eight hypotheses were supported: Perceived Usefulness exhibited the strongest effect on Behavioral Intention (β\u0026thinsp;=\u0026thinsp;0.312, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), followed by Trust (β\u0026thinsp;=\u0026thinsp;0.203, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), Perceived Learning Value and Perceived Ease of Use (both β\u0026thinsp;=\u0026thinsp;0.187, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), Perceived Risk (β = -0.156, p\u0026thinsp;=\u0026thinsp;0.001), AI Literacy (β\u0026thinsp;=\u0026thinsp;0.134, p\u0026thinsp;=\u0026thinsp;0.006), and Hedonic Motivation (β\u0026thinsp;=\u0026thinsp;0.098, p\u0026thinsp;=\u0026thinsp;0.045). Notably, H5 (Anxiety \u0026rarr; BI) was not supported (β\u0026thinsp;=\u0026thinsp;0.012, p\u0026thinsp;=\u0026thinsp;0.774). For moderation effects, four of six hypotheses were supported: Gender significantly moderated PEOU\u0026rarr;BI (H9a) and ANX\u0026rarr;BI (H9c), while IT Experience moderated PU\u0026rarr;BI (H10a) and AIL\u0026rarr;BI (H10b). H9b (Gender \u0026times; PR) and H10c (ITE \u0026times; TR) were not supported.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec31\" class=\"Section2\"\u003e \u003ch2\u003e4.3 Moderation Analysis\u003c/h2\u003e \u003cp\u003eMulti-group analysis revealed significant moderation effects for four of six hypotheses. Figure\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e illustrates gender moderation effects, and Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e presents IT experience moderation effects.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e presents the gender moderation analysis through two complementary visualizations. Panel (a) displays an interaction plot for the Perceived Ease of Use \u0026rarr; Behavioral Intention relationship, with PEOU scores on the x-axis (low to high) and Behavioral Intention on the y-axis. Two lines represent male and female subgroups, clearly demonstrating that female students (steeper slope, β\u0026thinsp;=\u0026thinsp;0.279) exhibit a substantially stronger positive relationship between ease of use perceptions and adoption intention compared to male students (shallower slope, β\u0026thinsp;=\u0026thinsp;0.112). The diverging slopes illustrate that as perceived ease of use increases, females' intention to adopt ChatGPT increases more dramatically than males'. Panel (b) presents a multi-group comparison bar chart showing path coefficient differences (Δβ) across all tested paths. Bars extending beyond the significance threshold lines indicate statistically significant moderation effects. Two paths show significant gender differences: PEOU\u0026rarr;BI (Δβ = -0.167, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001) and ANX\u0026rarr;BI (Δβ\u0026thinsp;=\u0026thinsp;0.109, p\u0026thinsp;=\u0026thinsp;0.033), while PR\u0026rarr;BI shows no significant difference (p\u0026thinsp;=\u0026thinsp;0.411). This visualization confirms that gender moderates effort expectancy and anxiety effects but not risk perception effects on ChatGPT adoption intention.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab7\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 7\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMulti-Group Analysis \u0026ndash; Gender Moderation\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePath\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ (Male)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ (Female)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eΔβ\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003et-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePEOU \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.112\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.279\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.167\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e3.345\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePR \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.134\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.178\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.044\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.823\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.411\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eANX \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.089\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e-0.198\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.109\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.134\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.033*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e\u003cem\u003eNote: ***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001, *p\u0026thinsp;\u0026lt;\u0026thinsp;0.05. Δβ\u0026thinsp;=\u0026thinsp;difference in path coefficients (Male \u0026ndash; Female).\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e presents the multi-group analysis results examining gender as a moderator of key structural relationships. Path coefficients are reported separately for male (n\u0026thinsp;=\u0026thinsp;254) and female (n\u0026thinsp;=\u0026thinsp;232) subgroups, along with coefficient differences (Δβ), t-values, and significance levels based on permutation testing with 5,000 permutations. Two paths exhibited significant gender differences: First, the Perceived Ease of Use \u0026rarr; Behavioral Intention relationship was significantly stronger for female students (β\u0026thinsp;=\u0026thinsp;0.279) compared to male students (β\u0026thinsp;=\u0026thinsp;0.112), with a significant difference of Δβ = -0.167 (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), supporting H9a and aligning with UTAUT predictions that effort expectancy effects are stronger for women [13]. Second, the Anxiety \u0026rarr; Behavioral Intention path showed a stronger negative effect for females (β = -0.198) than males (β = -0.089), with Δβ\u0026thinsp;=\u0026thinsp;0.109 (p\u0026thinsp;=\u0026thinsp;0.033), supporting H9c. The Perceived Risk \u0026rarr; Behavioral Intention path did not differ significantly across genders (p\u0026thinsp;=\u0026thinsp;0.411), failing to support H9b and suggesting risk concerns affect both genders similarly in programming education contexts.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e presents the IT experience moderation analysis through two complementary visualizations. Panel (a) displays an interaction plot for the Perceived Usefulness \u0026rarr; Behavioral Intention relationship, with PU scores on the x-axis and Behavioral Intention on the y-axis. Two lines represent high IT experience (n\u0026thinsp;=\u0026thinsp;187) and low-moderate IT experience (n\u0026thinsp;=\u0026thinsp;299) subgroups. The plot demonstrates that students with high IT experience (steeper slope, β\u0026thinsp;=\u0026thinsp;0.412) exhibit a substantially stronger positive relationship between usefulness perceptions and adoption intention compared to low-experience students (shallower slope, β\u0026thinsp;=\u0026thinsp;0.234). This indicates that experienced users are better able to recognize and weight performance benefits when making adoption decisions. Panel (b) presents a multi-group comparison bar chart showing path coefficient differences across all tested paths. Three paths show significant differences: PU\u0026rarr;BI (Δβ\u0026thinsp;=\u0026thinsp;0.178, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), AIL\u0026rarr;BI (Δβ\u0026thinsp;=\u0026thinsp;0.120, p\u0026thinsp;=\u0026thinsp;0.019), and an exploratory finding for PEOU\u0026rarr;BI (Δβ = -0.111, p\u0026thinsp;=\u0026thinsp;0.034) indicating that ease of use matters more for novice users. The TR\u0026rarr;BI path shows no significant difference (p\u0026thinsp;=\u0026thinsp;0.405), suggesting trust affects adoption similarly regardless of experience level.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab8\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 8\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMulti-Group Analysis \u0026ndash; IT Experience Moderation\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePath\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eβ (High)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eβ (Low)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eΔβ\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003et-value\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ep-value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePU \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.412\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.234\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.178\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e3.567\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAIL \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.198\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.078\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.120\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.345\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.019*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTR \u0026rarr; BI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.223\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.178\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.045\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.834\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.405\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePEOU \u0026rarr; BI\u0026dagger;\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.134\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.245\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e-0.111\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e2.123\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.034*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e\u003cem\u003eNote: ***p\u0026thinsp;\u0026lt;\u0026thinsp;0.001, *p\u0026thinsp;\u0026lt;\u0026thinsp;0.05. \u0026dagger;Exploratory finding (not hypothesized). High\u0026thinsp;=\u0026thinsp;high IT experience; Low\u0026thinsp;=\u0026thinsp;low-moderate IT experience.\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab8\" class=\"InternalRef\"\u003e8\u003c/span\u003e presents the multi-group analysis results examining IT experience as a moderator of structural relationships. Path coefficients are reported for high IT experience (n\u0026thinsp;=\u0026thinsp;187) and low-moderate IT experience (n\u0026thinsp;=\u0026thinsp;299) subgroups. Three significant differences emerged: First, the Perceived Usefulness \u0026rarr; Behavioral Intention relationship was substantially stronger for high-experience students (β\u0026thinsp;=\u0026thinsp;0.412) compared to low-experience students (β\u0026thinsp;=\u0026thinsp;0.234), with Δβ\u0026thinsp;=\u0026thinsp;0.178 (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), supporting H10a and indicating that experienced users place greater weight on performance benefits [13, 24]. Second, the AI Literacy \u0026rarr; Behavioral Intention effect was stronger for high-experience students (β\u0026thinsp;=\u0026thinsp;0.198 vs. β\u0026thinsp;=\u0026thinsp;0.078, p\u0026thinsp;=\u0026thinsp;0.019), supporting H10b and suggesting that technical background amplifies the benefits of AI knowledge [21, 25]. Third, an exploratory finding revealed that PEOU\u0026rarr;BI was stronger for low-experience students (β\u0026thinsp;=\u0026thinsp;0.245 vs. β\u0026thinsp;=\u0026thinsp;0.134, p\u0026thinsp;=\u0026thinsp;0.034), indicating usability concerns are more salient for novice users. The Trust \u0026rarr; Behavioral Intention path did not differ significantly (p\u0026thinsp;=\u0026thinsp;0.405), failing to support H10c.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec32\" class=\"Section2\"\u003e \u003ch2\u003e4.4 Model Fit, Effect Sizes, and Practical Significance\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab9\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 9\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eModel Fit and Predictive Validity\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIndex\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eValue\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eThreshold\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAssessment\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSRMR\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.052\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGood fit\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eR\u0026sup2; (Behavioral Intention)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.714\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.26 (substantial)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSubstantial\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eQ\u0026sup2; (Behavioral Intention)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.548\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.35 (large)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLarge\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGoF (Goodness of Fit)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.634\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u0026gt;\u0026thinsp;0.36 (large)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLarge\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab9\" class=\"InternalRef\"\u003e9\u003c/span\u003e presents the model fit and predictive validity indices for the AdoptGPT-Prog structural model. Four key indicators were evaluated: The Standardized Root Mean Square Residual (SRMR) value of 0.052 falls well below the 0.08 threshold, indicating good overall model fit with minimal discrepancy between observed and predicted correlations. The coefficient of determination (R\u0026sup2;) for Behavioral Intention was 0.714, substantially exceeding the 0.26 threshold for substantial explanatory power and indicating that the model explains 71.4% of variance in students' intention to adopt ChatGPT for programming learning. The Stone-Geisser Q\u0026sup2; value of 0.548 exceeds the 0.35 threshold for large predictive relevance, confirming the model's capability to accurately predict out-of-sample data. The Goodness-of-Fit (GoF) index of 0.634 surpasses the 0.36 threshold for large effect, representing the geometric mean of average communality and R\u0026sup2; values. Collectively, these indices confirm that AdoptGPT-Prog demonstrates excellent model fit, substantial explanatory power, and strong predictive validity.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e presents the Importance-Performance Map Analysis (IPMA) for Behavioral Intention, plotting construct importance (total effects, x-axis) against construct performance (mean rescaled scores 0-100, y-axis). This matrix identifies strategic priorities by positioning constructs in four quadrants. The upper-right quadrant (high importance, high performance) contains Perceived Usefulness and Trust, indicating these are key strengths that should be maintained\u0026mdash;students already perceive ChatGPT as useful and trustworthy, and these factors strongly drive adoption. The upper-left quadrant (low importance, high performance) contains Hedonic Motivation, suggesting enjoyment is not a critical adoption driver despite adequate performance. Perceived Ease of Use and Perceived Learning Value appear in the middle-right area with moderate-high importance and performance, representing factors to monitor and potentially enhance. The lower-right quadrant (high importance, low performance) identifies improvement priorities\u0026mdash;Perceived Risk and Anxiety appear here as barriers with meaningful negative effects and suboptimal scores, indicating that addressing misinformation concerns and AI-related apprehension could yield adoption benefits. AI Literacy shows moderate importance with room for performance improvement, suggesting AI education initiatives could strengthen adoption. This IPMA guides practitioners to prioritize maintaining usefulness perceptions while actively mitigating risk and anxiety concerns.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab10\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 10\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eEffect Sizes (f\u0026sup2;) and Practical Significance\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePredictor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ef\u0026sup2;\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEffect Size\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePractical Implication\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerceived Usefulness (PU)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.145\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMedium\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePrimary driver; high priority for intervention\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTrust (TR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.062\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSmall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eMeaningful impact; address reliability concerns\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerceived Ease of Use (PEOU)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.052\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSmall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eImportant for females and novice users\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerceived Learning Value (PLV)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.048\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSmall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eEmphasize educational benefits\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerceived Risk (PR)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSmall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eAddress misinformation and integrity concerns\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAI Literacy (AIL)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.027\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSmall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eGreater impact for experienced users\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHedonic Motivation (HM)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.015\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSmall\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLimited practical impact in this context\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAnxiety (ANX)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.002\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNegligible\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNot a significant barrier in this sample\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003cem\u003eNote: f\u0026sup2; thresholds: 0.02\u0026thinsp;=\u0026thinsp;small, 0.15\u0026thinsp;=\u0026thinsp;medium, 0.35\u0026thinsp;=\u0026thinsp;large [Cohen, 1988].\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab10\" class=\"InternalRef\"\u003e10\u003c/span\u003e reports the effect sizes (f\u0026sup2;) for each predictor's unique contribution to explaining Behavioral Intention, along with practical significance interpretations following Cohen's (1988) guidelines (small\u0026thinsp;\u0026ge;\u0026thinsp;0.02, medium\u0026thinsp;\u0026ge;\u0026thinsp;0.15, large\u0026thinsp;\u0026ge;\u0026thinsp;0.35). Perceived Usefulness demonstrated the largest effect (f\u0026sup2; = 0.145), approaching medium-level practical significance and establishing it as the primary driver warranting intervention priority. Trust (f\u0026sup2; = 0.062), Perceived Ease of Use (f\u0026sup2; = 0.052), Perceived Learning Value (f\u0026sup2; = 0.048), and Perceived Risk (f\u0026sup2; = 0.038) showed small but meaningful effects, each contributing incrementally to explained variance. AI Literacy (f\u0026sup2; = 0.027) and Hedonic Motivation (f\u0026sup2; = 0.015) exhibited smaller effects, with AI Literacy showing greater impact specifically for experienced users. Anxiety demonstrated negligible effect (f\u0026sup2; = 0.002), consistent with its non-significant path coefficient and indicating it is not a meaningful barrier in this programming student sample. These effect sizes inform practical prioritization: interventions should primarily target usefulness perceptions and trust-building, with secondary attention to ease of use (especially for females and novices) and risk mitigation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure \u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e presents a horizontal bar chart comparing effect sizes (f\u0026sup2;) of all eight predictors on Behavioral Intention, with vertical reference lines indicating Cohen's (1988) thresholds for small (0.02), medium (0.15), and large (0.35) effects. Bars are color-coded to distinguish enablers (blue) from barriers (red). Perceived Usefulness exhibits the largest effect (f\u0026sup2; = 0.145), approaching the medium effect threshold and visually dominating other predictors, confirming its role as the primary adoption driver. Trust (f\u0026sup2; = 0.062), Perceived Ease of Use (f\u0026sup2; = 0.052), Perceived Learning Value (f\u0026sup2; = 0.048), and Perceived Risk (f\u0026sup2; = 0.038) cluster in the small effect range, each making meaningful but more modest unique contributions. AI Literacy (f\u0026sup2; = 0.027) and Hedonic Motivation (f\u0026sup2; = 0.015) show smaller effects near the small effect threshold. Anxiety displays a negligible effect (f\u0026sup2; = 0.002), visually apparent as barely extending from the y-axis, consistent with its non-significant path coefficient. This visualization immediately communicates the practical significance hierarchy: usefulness perceptions deserve primary intervention focus, followed by trust-building and ease-of-use enhancements, while anxiety reduction may not yield substantial adoption benefits in programming student populations.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec33\" class=\"Section2\"\u003e \u003ch2\u003e4.5 Comparison with Prior Studies\u003c/h2\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab11\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 11\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eComparison with Recent ChatGPT Adoption Studies\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStudy\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eModel\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eN\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eR\u0026sup2; (BI)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eContext\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eModerators\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBalaskas et al. [8]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTAM+\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e423\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGeneral HE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eTrust, Risk (mediators)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRaman et al. [9]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDOI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e198\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGeneral HE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGender\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAl-Sharafi et al. [17]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003emeta-UTAUT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e355\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGeneral HE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePolicy\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSun et al. [4]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTAM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e82\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eProgramming\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNone\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCabero-Almenara et al. [32]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eUTAUT2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e612\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.64\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eGeneral HE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNone\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eThis Study\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTAM\u0026thinsp;+\u0026thinsp;UTAUT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e486\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.714\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eProgramming\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eGender, IT Experience\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003e\u003cem\u003eNote: HE\u0026thinsp;=\u0026thinsp;Higher Education. R\u0026sup2; values indicate variance explained in Behavioral Intention.\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab11\" class=\"InternalRef\"\u003e11\u003c/span\u003e provides a comparative analysis of AdoptGPT-Prog against five recent ChatGPT adoption studies in higher education contexts. Comparisons are made across theoretical model, sample size, explained variance (R\u0026sup2;), educational context, and moderating variables examined. AdoptGPT-Prog achieves the highest explanatory power (R\u0026sup2; = 0.714) among all compared studies, outperforming Balaskas et al.'s [8] TAM-based model (R\u0026sup2; = 0.58), Raman et al.'s [9] DOI model (R\u0026sup2; = 0.51), Al-Sharafi et al.'s [17] meta-UTAUT model (R\u0026sup2; = 0.62), Sun et al.'s [4] programming-focused TAM (R\u0026sup2; = 0.47), and Cabero-Almenara et al.'s [32] UTAUT2 model (R\u0026sup2; = 0.64). Notably, this study is the only one combining programming-specific context with systematic examination of both gender and IT experience as moderators. While Sun et al. [4] addressed programming education, their smaller sample (n\u0026thinsp;=\u0026thinsp;82) and absence of moderators limited generalizability. The integrated TAM-UTAUT framework employed here, combining eight predictors with dual moderators, provides more comprehensive theoretical coverage and superior predictive capability for understanding ChatGPT adoption specifically in programming education.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec34\" class=\"Section2\"\u003e \u003ch2\u003e4.6 Summary: Unsupported Hypotheses and Explanations\u003c/h2\u003e \u003cp\u003eThree hypotheses were not supported, warranting detailed explanation:\u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH5 (ANX \u0026rarr; BI) Not Supported\u003c/strong\u003e \u003cp\u003eContrary to general technology acceptance findings [14], anxiety did not significantly predict adoption intention (β\u0026thinsp;=\u0026thinsp;0.012, p\u0026thinsp;=\u0026thinsp;0.774, f\u0026sup2; = 0.002). This may reflect the unique characteristics of programming students who have self-selected into technical disciplines and possess baseline comfort with technology. Unlike general populations, these students may have overcome initial technology apprehension through their educational trajectory. Additionally, the framing of ChatGPT as a learning aid rather than an evaluative tool may reduce anxiety-inducing conditions. This finding aligns with Zhang and Yu [35], who found that familiarity with AI systems diminishes anxiety effects over time.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH9b (Gender \u0026times; PR \u0026rarr; BI) Not Supported\u003c/strong\u003e \u003cp\u003eThe borderline non-significant result (p\u0026thinsp;=\u0026thinsp;0.079) suggests that while a trend exists toward gender differences in risk perception effects, the programming education context may equalize risk concerns across genders. Both male and female programming students face similar concerns about code accuracy, academic integrity, and over-reliance, potentially diminishing gender-based differences observed in general technology contexts [10, 11].\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eH10c (ITE \u0026times; TR \u0026rarr; BI) Not Supported\u003c/strong\u003e \u003cp\u003eTrust affected adoption similarly across experience levels (p\u0026thinsp;=\u0026thinsp;0.125), contradicting our expectation that experienced users would rely less on trust. This suggests that trust in ChatGPT's accuracy and reliability remains universally important regardless of technical background\u0026mdash;a finding consistent with Niu and Mvondo [18], who argued that trust serves as a foundational prerequisite that does not diminish with experience when adopting novel AI tools.\u003c/p\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"5. Discussion","content":"\u003cp\u003eThe AdoptGPT-Prog model explained 71.4% of variance in Behavioral Intention, substantially outperforming prior ChatGPT adoption studies (R\u0026sup2; range: 0.47\u0026ndash;0.64) [4, 8, 17, 32]. This section interprets key findings in relation to existing literature and addresses unsupported hypotheses.\u003c/p\u003e \u003cdiv id=\"Sec36\" class=\"Section2\"\u003e \u003ch2\u003e5.1 Perceived Usefulness as the Dominant Predictor\u003c/h2\u003e \u003cp\u003ePerceived Usefulness emerged as the strongest predictor (β\u0026thinsp;=\u0026thinsp;0.312, f\u0026sup2; = 0.145), confirming TAM's core proposition [7] and aligning with Zhang and Pan [34], who found performance expectancy significantly predicted Duolingo adoption. Students who recognize ChatGPT's value for debugging, code generation, and concept explanation demonstrate substantially higher adoption intention. This consistency across AI-assisted programming (our study) and gamified language learning [34] suggests performance benefits serve as universal adoption drivers in educational technology.\u003c/p\u003e \u003cp\u003eZhang and Yu [35] similarly found that performance expectancy directly predicted continuance intention for the DouBao chatbot among EFL learners, with emotional attachment and trust as mediators. The parallel demonstrates that perceived performance benefits operate consistently across generative AI applications in education.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec37\" class=\"Section2\"\u003e \u003ch2\u003e5.2 The Critical Role of Trust\u003c/h2\u003e \u003cp\u003eTrust showed the second-strongest effect (β\u0026thinsp;=\u0026thinsp;0.203, f\u0026sup2; = 0.062), consistent with Al-Sharafi et al. [17] and Niu and Mvondo [18], who established trust as a cornerstone of ChatGPT adoption in higher education. Zhang and Yu [35] found that trust mediated the relationship between perceived anthropomorphism and performance expectancy, suggesting trust serves as a critical intermediary through which users evaluate AI capabilities. Our finding that trust affected adoption similarly across IT experience levels (H10c not supported) extends this by indicating trust remains universally important regardless of technical background\u0026mdash;a foundational prerequisite that does not diminish with experience [18].\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec38\" class=\"Section2\"\u003e \u003ch2\u003e5.3 Perceived Ease of Use and Learning Value\u003c/h2\u003e \u003cp\u003eBoth PEOU (β\u0026thinsp;=\u0026thinsp;0.187) and PLV (β\u0026thinsp;=\u0026thinsp;0.187) significantly predicted adoption intention. Deng and Yu [36] found that perceived ease of use influenced curiosity, control, and joy in TikTok adoption, confirming ease of use as a foundational enabler unlocking cognitive and emotional pathways. The gender moderation effect\u0026mdash;where females showed stronger PEOU\u0026rarr;BI effects (β\u0026thinsp;=\u0026thinsp;0.279 vs. 0.112)\u0026mdash;aligns with UTAUT predictions [13] and suggests female programming students may be more sensitive to interface design when navigating technically complex AI tools.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec39\" class=\"Section2\"\u003e \u003ch2\u003e5.4 Risk Perceptions and the Non-Significant Anxiety Effect\u003c/h2\u003e \u003cp\u003ePerceived Risk significantly inhibited adoption (β = -0.156), confirming that concerns about misinformation, academic integrity, and privacy deter students [6, 8]. However, Anxiety did not significantly predict Behavioral Intention (H5: β\u0026thinsp;=\u0026thinsp;0.012, p\u0026thinsp;=\u0026thinsp;0.774, f\u0026sup2; = 0.002)\u0026mdash;a notable divergence from general technology acceptance findings [14].\u003c/p\u003e \u003cp\u003eThis null finding may reflect population characteristics: programming students have self-selected into technical disciplines and likely possess baseline comfort with technology, having overcome initial apprehension through prior computational exposure. Additionally, ChatGPT's framing as a learning aid rather than evaluative tool may reduce anxiety-inducing conditions. Zhang and Yu [35] noted that emotional factors in AI contexts operate through complex mediating mechanisms rather than simple direct effects, which may explain why anxiety's influence is attenuated when trust and usefulness are controlled. Notably, the gender moderation analysis revealed anxiety significantly affects females more strongly (β = -0.198 vs. -0.089, p\u0026thinsp;=\u0026thinsp;0.033), indicating anxiety operates as a conditional barrier whose influence varies by demographic subgroup.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec40\" class=\"Section2\"\u003e \u003ch2\u003e5.5 AI Literacy and IT Experience Moderation\u003c/h2\u003e \u003cp\u003eAI Literacy positively predicted adoption (β\u0026thinsp;=\u0026thinsp;0.134), with IT experience significantly moderating this relationship (High IT: β\u0026thinsp;=\u0026thinsp;0.198 vs. Low IT: β\u0026thinsp;=\u0026thinsp;0.078). Students with higher IT experience possess cognitive frameworks that amplify AI literacy benefits, enabling more effective translation of AI knowledge into usage strategies [21, 25]. This parallels Zhang and Pan's [34] finding that environmental factors enhancing user competencies facilitate adoption, and Deng and Yu's [36] finding that personal innovativeness influenced ease of use perceptions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec41\" class=\"Section2\"\u003e \u003ch2\u003e5.6 Hedonic Motivation: Secondary but Significant\u003c/h2\u003e \u003cp\u003eHedonic Motivation showed a modest but significant effect (β\u0026thinsp;=\u0026thinsp;0.098, f\u0026sup2; = 0.015), contrasting with Deng and Yu's [36] TikTok study where curiosity was a major predictor (β\u0026thinsp;=\u0026thinsp;0.395). This difference reflects the fundamental distinction between hedonic-oriented systems (entertainment platforms) and utilitarian-oriented systems (learning tools). In programming education, students prioritize practical outcomes over enjoyment, explaining why hedonic factors serve as complementary rather than primary drivers\u0026mdash;consistent with Zhang and Pan's [34] finding that hedonic motivation had no significant effect on Duolingo adoption when utilitarian learning goals predominated.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec42\" class=\"Section2\"\u003e \u003ch2\u003e5.7 Demographic Moderation Effects\u003c/h2\u003e \u003cp\u003eGender moderated PEOU\u0026rarr;BI and ANX\u0026rarr;BI relationships, while IT experience moderated PU\u0026rarr;BI and AIL\u0026rarr;BI. These findings contrast with Deng and Yu [36], who found no gender moderation in TikTok adoption, attributed to universal platform design. Our significant gender effects likely reflect the technical nature of programming and persistent STEM gender disparities\u0026mdash;female students may face additional barriers related to confidence and stereotype threat, making usability and anxiety reduction particularly important [10, 22]. The IT experience moderation of PU\u0026rarr;BI (High: β\u0026thinsp;=\u0026thinsp;0.412 vs. Low: β\u0026thinsp;=\u0026thinsp;0.234) parallels findings that experienced users develop more sophisticated evaluation frameworks [13, 24], enabling accurate assessment of how ChatGPT capabilities translate to performance improvements.\u003c/p\u003e \u003c/div\u003e"},{"header":"6. Implications","content":"\u003cdiv id=\"Sec44\" class=\"Section2\"\u003e \u003ch2\u003e6.1 Theoretical Implications\u003c/h2\u003e \u003cp\u003eThis study makes three theoretical contributions: (1) It extends TAM-UTAUT frameworks by integrating domain-specific constructs (AI Literacy, Perceived Learning Value) for generative AI contexts, achieving superior explanatory power (R\u0026sup2; = 0.714) compared to existing models [4, 8, 17, 32]. (2) It demonstrates that demographic moderators systematically influence adoption pathways, supporting the need for differentiated acceptance models rather than universal frameworks. (3) The non-significant anxiety finding challenges assumptions from general technology acceptance research, suggesting domain-specific populations (e.g., programming students) may exhibit distinct adoption patterns requiring context-sensitive theorizing.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec45\" class=\"Section2\"\u003e \u003ch2\u003e6.2 Practical Implications\u003c/h2\u003e \u003cp\u003eBased on effect size analysis (Table\u0026nbsp;\u003cspan refid=\"Tab10\" class=\"InternalRef\"\u003e10\u003c/span\u003e) and moderation findings, we offer stratified recommendations in Table\u0026nbsp;\u003cspan refid=\"Tab12\" class=\"InternalRef\"\u003e12\u003c/span\u003e:\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab12\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 12\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eStratified Practical Recommendations\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"3\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePriority\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTarget Factor\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSpecific Recommendations\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHIGH\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePerceived Usefulness (f\u0026sup2; = 0.145)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDemonstrate ChatGPT's value for debugging, code explanation, and productivity through hands-on workshops; showcase before/after performance comparisons\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHIGH\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eTrust (f\u0026sup2; = 0.062)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eProvide guided demonstrations showing output accuracy; establish clear guidelines on when to trust vs. verify AI responses; address hallucination concerns explicitly\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMEDIUM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePEOU \u0026ndash; Females/Novices\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDesign intuitive onboarding for first-time users; provide scaffolded prompting templates; offer additional support for female students navigating technical interfaces\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMEDIUM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePerceived Risk (f\u0026sup2; = 0.038)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDevelop institutional AI use policies addressing academic integrity; teach critical evaluation of AI-generated code; clarify privacy implications\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMEDIUM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAI Literacy \u0026ndash; High IT Exp.\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIntegrate AI literacy modules into programming curricula; teach effective prompting strategies; explain AI limitations and appropriate use cases\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLOW\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAnxiety \u0026ndash; Female Students\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCreate supportive, low-stakes environments for AI experimentation; normalize uncertainty in AI interactions; provide reassurance about learning curves\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLOW\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eHedonic Motivation (f\u0026sup2; = 0.015)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eWhile not primary, engaging interactions can complement performance focus; consider gamified prompting challenges for interested students\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"3\"\u003eNote: Priority based on effect sizes and moderation significance. HIGH\u0026thinsp;=\u0026thinsp;primary intervention targets; MEDIUM\u0026thinsp;=\u0026thinsp;secondary considerations; LOW\u0026thinsp;=\u0026thinsp;complementary factors.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"7. Limitations and Future Research","content":"\u003cp\u003e \u003cstrong\u003eLimitations\u003c/strong\u003e \u003cp\u003e(1) Cross-sectional design captures perceptions at one time point; longitudinal studies could examine how perceptions evolve with sustained ChatGPT use. (2) The sample is limited to programming students at Saudi Arabian universities, constraining generalizability to other disciplines and cultural contexts. (3) Self-reported behavioral intention may not fully predict actual usage behavior (intention-behavior gap). (4) The non-significant anxiety finding, while explained theoretically, may require replication in other technical student populations. (5) Potential construct overlap between PU and PLV (HTMT\u0026thinsp;=\u0026thinsp;0.723), though within acceptable bounds, warrants attention in future model refinements.\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eFuture Research Directions\u003c/strong\u003e \u003cp\u003e(1) Longitudinal studies tracking adoption patterns and skill development over academic terms. (2) Replication across disciplines (e.g., data science, web development) and cultural contexts. (3) Integration of objective usage metrics (e.g., ChatGPT interaction logs) to validate self-reported intentions. (4) Examination of additional constructs such as self-efficacy, cognitive load, and ethical awareness. (5) Comparative studies between different generative AI tools (e.g., ChatGPT vs. GitHub Copilot vs. Claude) in programming education contexts.\u003c/p\u003e \u003c/p\u003e"},{"header":"8. Conclusion","content":"\u003cp\u003eThis study developed and validated AdoptGPT-Prog, an integrated TAM-UTAUT model for understanding ChatGPT adoption in programming education. Based on 486 undergraduate students from five Saudi Arabian universities, the model achieved substantial explanatory power (R\u0026sup2; = 0.714), outperforming prior ChatGPT adoption frameworks.\u003c/p\u003e \u003cp\u003eKey findings reveal that Perceived Usefulness (β\u0026thinsp;=\u0026thinsp;0.312) and Trust (β\u0026thinsp;=\u0026thinsp;0.203) are the primary adoption drivers, while Perceived Risk (β = -0.156) serves as a significant barrier. Notably, Anxiety did not significantly predict adoption intention in this programming student sample\u0026mdash;a context-specific finding suggesting technical students may have overcome technology apprehension through prior exposure. Gender moderates the effects of Perceived Ease of Use and Anxiety on intention, with female students showing stronger sensitivity to usability factors. IT Experience moderates the effects of Perceived Usefulness and AI Literacy, with experienced students placing greater weight on performance benefits.\u003c/p\u003e \u003cp\u003eThese results provide actionable guidance for educators: prioritize demonstrating ChatGPT's practical value and building trust through guided practice, while addressing risk concerns through clear institutional policies and AI literacy education. Differentiated approaches should accommodate gender and experience differences in adoption patterns. The AdoptGPT-Prog model contributes an empirically validated, context-sensitive framework that advances understanding of generative AI acceptance in programming education and informs evidence-based integration strategies.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv align=\"left\"\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"376\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePEOU\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePerceived Ease of Use\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003ePerceived Risk\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAIL\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eAI Literacy\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eTrust\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eANX\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eAnxiety\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePU\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003ePerceived Usefulness\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePLV\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003ePerceived Learning Value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHM\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eHedonic Motivation\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eBI\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eBehavioral Intention\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGEN\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eGender (moderator)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eITE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eIT Experience (moderator)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cimg width=\"8\" height=\"18\" src=\"data:image/png;base64,R0lGODlhDAAbAHcAMSH+GlNvZnR3YXJlOiBNaWNyb3NvZnQgT2ZmaWNlACH5BAEAAAAALAEACQALABIAhAAAAAAAAAAAOgAAZgA6OgA6ZgA6kABmtjoAADpmtjqQ22YAAGY6AGaQ22a225A6AJC225Db/7ZmALZmOrbb/7b//9uQOtv///+2Zv/bkP/btv//tv//2wECAwECAwECAwVYIABwExIERCMCGxM4gHUqohQYF7vc17YfIh8vY6IBiAEgZgbQPHiAZUAhTeSiTGdAEMFOgzuc1AiwcccrcwQ9eg4q0gElxpQSTIECRMRe8Zl+f1+BXmR+IQA7\" alt=\"image\"\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eStandardized path coefficient\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cimg width=\"7\" height=\"18\" src=\"data:image/png;base64,R0lGODlhCwAbAHcAMSH+GlNvZnR3YXJlOiBNaWNyb3NvZnQgT2ZmaWNlACH5BAEAAAAALAAACQALAA4AhAAAAAAAAAAAOgAAZgA6kABmtjoAADo6ADpmtjqQ22Y6AGaQ22a2/5A6AJC225Db/7ZmALZmZrb//9uQOtuQZtv///+2Zv/bkP//2wECAwECAwECAwECAwECAwECAwECAwVEIABcSoCI6GgEgfCkQOQAUJDA6GUQFQ5gjYHEp3P5agEGTse6pYCtQwFmYRUmPBTUZRGiJiyeTqliKYE8CtILsBgEixAAOw==\" alt=\"image\"\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eFactor loading\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cimg width=\"15\" height=\"19\" src=\"data:image/png;base64,R0lGODlhFwAcAHcAMSH+GlNvZnR3YXJlOiBNaWNyb3NvZnQgT2ZmaWNlACH5BAEAAAAALAAABwAWABEAhAAAAAAAAAAAOgAAZgA6ZgA6kABmtjoAADo6ADo6ZjpmkDpmtjqQ22YAAGY6OmaQ22a2/5A6AJBmOpDb/7ZmALaQZrbb/7b//9uQOtv///+2Zv/bkP/btv//tv//2wECAwV/ICCOJOlVSBAQU+mSXcMAWFBkbz5ux3DpOk3AAIipjgGBwvKi3ESbRrLFkQpaJIxv5IlMRcIhaZP4cb1XsIoo6jhaHgkcjQ3PRBRkuvutBhY4ADxIW3xILEBnU35piXxpNTaBOo8tRgEQjnR4Kk86USoPUAcqdy+XKpkAeaU5IQA7\" alt=\"image\"\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eCoefficient of determination\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cimg width=\"15\" height=\"19\" src=\"data:image/png;base64,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\" alt=\"image\"\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003ePredictive relevance\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cimg width=\"12\" height=\"19\" src=\"data:image/png;base64,R0lGODlhEgAcAHcAMSH+GlNvZnR3YXJlOiBNaWNyb3NvZnQgT2ZmaWNlACH5BAEAAAAALAAABwARABEAhAAAAAAAAAAAOgAAZgA6ZgA6kABmtjoAADo6ADo6ZjqQ22YAAGY6OpBmOpDb/7ZmALaQZrbb/7b//9uQOtu2kNv///+2Zv/bkP//tv//2wECAwECAwECAwECAwECAwECAwVdICCOYwYhQUA4ZAtgiwJMQVG5+HUM0kgxKR7JEjCMaLyJcPSwiWC1W0vZE+mirUuiCiCmUsYng5VpsKBO5jcgOC+w1sNaiJbiRvU7Ka9/vtN9fH1eAxF9L28pgDghADs=\" alt=\"image\"\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eEffect size\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAVE\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eAverage Variance Extracted\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eCR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eComposite Reliability\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHTMT\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eHeterotrait-Monotrait ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSRMR\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eStandardized Root Mean Square Residual\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVIF\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eVariance Inflation Factor\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eGoF\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eGoodness-of-Fit Index\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003ePLS-SEM\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003ePartial Least Squares Structural Equation Modeling\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eTAM\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eTechnology Acceptance Model\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 26.5957%;\"\u003e\n \u003cp\u003e\u003cstrong\u003eUTAUT\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 73.4043%;\"\u003e\n \u003cp\u003eUnified Theory of Acceptance and Use of Technology\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor Contributions:\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFaisal Alshammari contributed to the conceptualization, methodology, software development, and provision of resources for the study. Both authors also collaborated on reviewing and editing the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgement:\u0026nbsp;\u003c/strong\u003eThe author extends the appreciation to the Deanship of Postgraduate Studies and Scientific Research at Majmaah University for funding this research work through the project number (XXXXXX).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding:\u003c/strong\u003e The author extends the appreciation to the Deanship of Postgraduate Studies and Scientific Research at Majmaah University for funding this research work through the project number (XXXXXX).\u003c/p\u003e\n\u003cp\u003eData Availability Statement:\u0026nbsp;Data supporting the findings of this study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of Interest:\u0026nbsp;\u003c/strong\u003eThe author declares no conflicts of interest regarding the publication of this research paper.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval:\u003c/strong\u003e All experimental procedures and clinical data collection were conducted in strict accordance with the principles outlined in the Declaration of Helsinki for medical research involving human subjects. The study protocol was re-viewed and approved by the Majmaah University for Research Ethics Committee (MUREC) (HA-01-R-088) under the Ethics number: \u003cstrong\u003eMUREC-\u003c/strong\u003e\u003cstrong\u003eDec\u003c/strong\u003e\u003cstrong\u003e.06/COM-2025/\u003c/strong\u003e\u003cstrong\u003e300\u003c/strong\u003e.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to publish:\u003c/strong\u003e Written informed consent was obtained from all participants prior to data collection, ensuring their voluntary participation and awareness of the study objectives.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical Trials:\u0026nbsp;\u003c/strong\u003eNot applicable\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eY. K. Dwivedi, N. Kshetri, L. Hughes, E. L. Slade, A. Jeyaraj, A. K. Kar, A. M. Baabdullah, A. Koohang, V. Raghavan, M. Ahuja, et al., \u0026ldquo;Opinion paper: So what if ChatGPT wrote it? Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy,\u0026rdquo; International Journal of Information Management, vol. 71, p. 102642, 2023. [Online]. Available: https://doi.org/10.1016/j.ijinfomgt.2023.102642\u003c/li\u003e\n \u003cli\u003eM. M. Rahman and Y. Watanobe, \u0026ldquo;ChatGPT for education and research: Opportunities, threats, and strategies,\u0026rdquo; Applied Sciences, vol. 13, no. 9, p. 5783, 2023. [Online]. Available: https://doi.org/10.3390/app13095783\u003c/li\u003e\n \u003cli\u003eE. Kasneci, K. Sessler, S. K\u0026uuml;chemann, M. Bannert, D. Dementieva, F. Fischer, U. Gasser, G. Groh, S. G\u0026uuml;nnemann, E. H\u0026uuml;llermeier, et al., \u0026ldquo;ChatGPT for good? On opportunities and challenges of large language models for education,\u0026rdquo; Learning and Individual Differences, vol. 103, p. 102274, 2023. [Online]. Available: https://doi.org/10.1016/j.lindif.2023.102274\u003c/li\u003e\n \u003cli\u003eD. Sun, Y. Bao, Y. Zhu, and Y. Chen, \u0026ldquo;Would ChatGPT-facilitated programming mode impact college students\u0026rsquo; programming behaviors, performances, and perceptions? An empirical study,\u0026rdquo; International Journal of Educational Technology in Higher Education, vol. 21, no. 14, 2024. [Online]. Available: https://doi.org/10.1186/s41239-024-00446-5\u003c/li\u003e\n \u003cli\u003eBiswas, \u0026ldquo;ChatGPT and the future of computer programming education,\u0026rdquo; Preprints, 2022. [Online]. Available: https://doi.org/10.20944/preprints202212.0175.v1\u003c/li\u003e\n \u003cli\u003eT. H. Baek and M. Kim, \u0026ldquo;Is ChatGPT scary good? How user motivations affect creepiness and trust in generative artificial intelligence,\u0026rdquo; Telematics and Informatics, vol. 83, p. 102030, 2023. [Online]. Available: https://doi.org/10.1016/j.tele.2023.102030\u003c/li\u003e\n \u003cli\u003eF. D. Davis, \u0026ldquo;Perceived usefulness, perceived ease of use, and user acceptance of information technology,\u0026rdquo; MIS Quarterly, vol. 13, no. 3, pp. 319\u0026ndash;340, 1989. [Online]. Available: https://doi.org/10.2307/249008\u003c/li\u003e\n \u003cli\u003eS. Balaskas, V. Tsiantos, S. Chatzifotiou, and M. Rigou, \u0026ldquo;Determinants of ChatGPT adoption intention in higher education: Expanding on TAM with the mediating roles of trust and risk,\u0026rdquo; Information, vol. 16, no. 2, p. 82, 2025. [Online]. Available: https://doi.org/10.3390/info16020082\u003c/li\u003e\n \u003cli\u003eR. Raman, K. Vinuesa, and S. Karthikeyan, \u0026ldquo;Exploring university students\u0026rsquo; adoption of ChatGPT using the diffusion of innovation theory and sentiment analysis with gender dimension,\u0026rdquo; Human Behavior and Emerging Technologies, vol. 2024, p. 3085910, 2024. [Online]. Available: https://doi.org/10.1155/2024/3085910\u003c/li\u003e\n \u003cli\u003eZ. Cai, X. Fan, and J. Du, \u0026ldquo;Gender and attitudes toward technology use: A meta-analysis,\u0026rdquo; Computers \u0026amp; Education, vol. 105, pp. 1\u0026ndash;13, 2017. [Online]. Available: https://doi.org/10.1016/j.compedu.2016.11.003\u003c/li\u003e\n \u003cli\u003eShaouf, K. Lu, and X. Li, \u0026ldquo;The impact of gender differences on adoption of information technology and related responses: A review,\u0026rdquo; International Journal of Management and Applied Research, vol. 5, no. 1, pp. 22\u0026ndash;41, 2018. [Online]. Available: https://doi.org/10.18646/2056.51.18-003\u003c/li\u003e\n \u003cli\u003eF. M. Yilmaz, A. A. Yilmaz, and S. Karakus, \u0026ldquo;The moderating effects of gender and study discipline in the relationship between university students\u0026rsquo; acceptance and use of ChatGPT,\u0026rdquo; European Journal of Investigation in Health, Psychology and Education, vol. 14, no. 7, pp. 1981\u0026ndash;1995, 2024. [Online]. Available: https://doi.org/10.3390/ejihpe14070132\u003c/li\u003e\n \u003cli\u003eV. Venkatesh, M. G. Morris, G. B. Davis, and F. D. Davis, \u0026ldquo;User acceptance of information technology: Toward a unified view,\u0026rdquo; MIS Quarterly, vol. 27, no. 3, pp. 425\u0026ndash;478, 2003. [Online]. Available: https://doi.org/10.2307/30036540\u003c/li\u003e\n \u003cli\u003eR. Scherer, F. Siddiq, and J. Tondeur, \u0026ldquo;The technology acceptance model (TAM): A meta-analytic structural equation modeling approach to explaining teachers\u0026rsquo; adoption of digital technology in education,\u0026rdquo; Computers \u0026amp; Education, vol. 128, pp. 13\u0026ndash;35, 2019. [Online]. Available: https://doi.org/10.1016/j.compedu.2018.09.009\u003c/li\u003e\n \u003cli\u003eM. A. Almaiah, R. Shishakly, and S. Al-Otaibi, \u0026ldquo;A new technological model on investigating the utilization of mobile learning applications: Extending the TAM,\u0026rdquo; Multimodal Technologies and Interaction, vol. 7, no. 9, p. 92, 2023. [Online]. Available: https://doi.org/10.3390/mti7090092\u003c/li\u003e\n \u003cli\u003eJ. M. Carvajal-Morales, E. E. Le\u0026oacute;n-Pl\u0026uacute;as, J. D. Valenzuela-Cobos, and F. Guevara-Viej\u0026oacute;, \u0026ldquo;Educational design in the adoption of ICT for sustainable digital learning in social and business sciences: A structural equation model,\u0026rdquo; Sustainability, vol. 16, no. 23, p. 10674, 2024. [Online]. Available: https://doi.org/10.3390/su162310674\u003c/li\u003e\n \u003cli\u003eM. A. Al-Sharafi, H. Al-Emran, I. Arpaci, N. Marques, M. F. Khani, and M. Ismail, \u0026ldquo;Understanding students\u0026rsquo; adoption of the ChatGPT chatbot in higher education: The role of anthropomorphism, trust, design novelty and institutional policy,\u0026rdquo; Behaviour \u0026amp; Information Technology, vol. 43, no. 2, pp. 315\u0026ndash;336, 2024. [Online]. Available: https://doi.org/10.1080/0144929X.2024.2317364\u003c/li\u003e\n \u003cli\u003eB. Niu and G. N. Mvondo, \u0026ldquo;ChatGPT awareness, acceptance, and adoption in higher education: The role of trust as a cornerstone,\u0026rdquo; International Journal of Educational Technology in Higher Education, vol. 21, no. 46, 2024. [Online]. Available: https://doi.org/10.1186/s41239-024-00478-x\u003c/li\u003e\n \u003cli\u003eL. J. Avila-Chauvet, D. G. Mej\u0026iacute;a, and C. D. Acosta Quiroz, \u0026ldquo;ChatGPT for programming assistance,\u0026rdquo; Behavioural Sciences, vol. 13, no. 8, p. 675, 2023. [Online]. Available: https://doi.org/10.3390/bs13080675\u003c/li\u003e\n \u003cli\u003eD. Xiao, Y. Bao, Y. Zhu, and D. Sun, \u0026ldquo;Investigating students\u0026rsquo; programming behaviors, interaction qualities and perceptions through prompt-based learning in ChatGPT,\u0026rdquo; Humanities and Social Sciences Communications, vol. 11, no. 1447, 2024. [Online]. Available: https://doi.org/10.1038/s41599-024-03991-6\u003c/li\u003e\n \u003cli\u003eX. Silva, Y. Chen, and Z. Wang, \u0026ldquo;What factors will affect the effectiveness of using ChatGPT to solve programming problems? A quasi-experimental study,\u0026rdquo; Humanities and Social Sciences Communications, vol. 11, no. 319, 2024. [Online]. Available: https://doi.org/10.1038/s41599-024-02751-w\u003c/li\u003e\n \u003cli\u003eC. Campos and R. Scherer, \u0026ldquo;Digital gender gaps in students\u0026rsquo; knowledge, attitudes and skills: An integrative data analysis across 32 countries,\u0026rdquo; Education and Information Technologies, vol. 29, pp. 4225\u0026ndash;4258, 2024. [Online]. Available: https://doi.org/10.1007/s10639-023-12272-9\u003c/li\u003e\n \u003cli\u003eY. Lu and S. Yang, \u0026ldquo;Adoption of multimedia technology for learning and gender difference,\u0026rdquo; Computers in Human Behavior, vol. 92, pp. 288\u0026ndash;296, 2019. [Online]. Available: https://doi.org/10.1016/j.chb.2018.11.029\u003c/li\u003e\n \u003cli\u003eD. M. Jacobsen, \u0026ldquo;Adoption patterns and characteristics of faculty who integrate computer technology for teaching and learning in higher education,\u0026rdquo; ERIC Document, ED418892, 1998.\u003c/li\u003e\n \u003cli\u003eE. Pesovski, R. Bogdanoski, and M. Trajkovik, \u0026ldquo;AI-based learning materials for programming education,\u0026rdquo; Education and Information Technologies, vol. 29, pp. 8945\u0026ndash;8970, 2024. [Online]. Available: https://doi.org/10.1007/s10639-023-12345-9\u003c/li\u003e\n \u003cli\u003eOkonkwo and A. Ade-Ibijola, \u0026ldquo;Chatbots applications in education: A systematic review,\u0026rdquo; Computers and Education: Artificial Intelligence, vol. 2, p. 100033, 2021. [Online]. Available: https://doi.org/10.1016/j.caeai.2021.100033\u003c/li\u003e\n \u003cli\u003eK. Malinka, M. Peres\u0026iacute;ni, A. Firc, O. Hujň\u0026aacute;k, and F. Janu\u0026scaron;, \u0026ldquo;On the educational impact of ChatGPT: Is artificial intelligence ready to obtain a university degree?\u0026rdquo; in Proc. ACM Conference on Innovation and Technology in Computer Science Education, 2023, pp. 47\u0026ndash;53. [Online]. Available: https://doi.org/10.1145/3587102.3588827\u003c/li\u003e\n \u003cli\u003eStrzelecki, \u0026ldquo;Students\u0026rsquo; acceptance of ChatGPT in higher education: An extended unified theory of acceptance and use of technology,\u0026rdquo; Innovative Higher Education, vol. 49, pp. 223\u0026ndash;245, 2024. [Online]. Available: https://doi.org/10.1007/s10755-023-09686-1\u003c/li\u003e\n \u003cli\u003eR. L. Thompson, C. A. Higgins, and J. M. Howell, \u0026ldquo;Personal computing: Toward a conceptual model of utilization,\u0026rdquo; MIS Quarterly, vol. 15, no. 1, pp. 125\u0026ndash;143, 1991. [Online]. Available: https://doi.org/10.2307/249443\u003c/li\u003e\n \u003cli\u003eS. Taylor and P. A. Todd, \u0026ldquo;Understanding information technology usage: A test of competing models,\u0026rdquo; Information Systems Research, vol. 6, no. 2, pp. 144\u0026ndash;176, 1995. [Online]. Available: https://doi.org/10.1287/isre.6.2.144\u003c/li\u003e\n \u003cli\u003eAjzen, \u0026ldquo;The theory of planned behavior,\u0026rdquo; Organizational Behavior and Human Decision Processes, vol. 50, no. 2, pp. 179\u0026ndash;211, 1991. [Online]. Available: https://doi.org/10.1016/0749-5978(91)90020-T\u003c/li\u003e\n \u003cli\u003eCabero-Almenara, A. Palacios-Rodr\u0026iacute;guez, H. A. Rojas Guzm\u0026aacute;n, and V. Fern\u0026aacute;ndez-Scagliusi, \u0026ldquo;Prediction of the use of generative artificial intelligence through ChatGPT among Costa Rican university students: A PLS model based on UTAUT2,\u0026rdquo; Applied Sciences, vol. 15, no. 6, p. 3363, 2025. [Online]. Available: https://doi.org/10.3390/app15063363\u003c/li\u003e\n \u003cli\u003eM. Al-Okaily, W. Mater, N. Matar, and F. S. Shiyyab, \u0026ldquo;The role of ChatGPT adoption in enhancing communication and collaboration skills among business school students,\u0026rdquo; Computers and Education: Artificial Intelligence, p. 100441, 2025. [Online]. Available: https://doi.org/10.1016/j.caeai.2025.100441\u003c/li\u003e\n \u003cli\u003eY. Zhang and W. Pan, \u0026quot;A UTAUT2 model expansion: Investigating the effect of interactive learning environment and gamification on Duolingo user base in China,\u0026quot; \u003cem\u003eInternational Journal of Human\u0026ndash;Computer Interaction\u003c/em\u003e, vol. 41, no. 15, pp. 9518\u0026ndash;9532, 2025. [Online]. Available: https://doi.org/10.1080/10447318.2024.2426736\u003c/li\u003e\n \u003cli\u003eY. Zhang and Z. Yu, \u0026quot;Emotional attachment as the key mediator: Expanding UTAUT2 to examine how perceived anthropomorphism in intelligent agents influences the sustained use of DouBao (Cici) among EFL learners,\u0026quot; \u003cem\u003eEducation and Information Technologies\u003c/em\u003e, vol. 30, pp. 23897\u0026ndash;23919, 2025. [Online]. Available: https://doi.org/10.1007/s10639-025-13721-3\u003c/li\u003e\n \u003cli\u003eX. Deng and Z. Yu, \u0026quot;An extended hedonic motivation adoption model of TikTok in higher education,\u0026quot; \u003cem\u003eEducation and Information Technologies\u003c/em\u003e, vol. 28, no. 10, pp. 13595\u0026ndash;13617, 2023. [Online]. Available: https://doi.org/10.1007/s10639-023-11749-x\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"discover-education","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"diedu","sideBox":"Learn more about [Discover Education](https://www.springer.com/journal/44217)","snPcode":"44217","submissionUrl":"https://submission.nature.com/new-submission/44217/3","title":"Discover Education","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"ChatGPT, Programming Education, Technology Acceptance Model, UTAUT, Gender Moderation, IT Experience, Behavioral Intention, PLS-SEM, Generative AI Adoption","lastPublishedDoi":"10.21203/rs.3.rs-8564643/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8564643/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eDespite growing interest in ChatGPT adoption in higher education, existing research lacks domain-specific models that account for the unique cognitive demands of programming education and the demographic heterogeneity among learners. This study addresses three research questions: (1) What is the relative importance of cognitive-affective factors (perceived usefulness, trust, anxiety, perceived risk) versus competency factors (AI literacy, perceived ease of use) in predicting programming students' ChatGPT adoption intention? (2) Do gender differences in technology adoption persist when examining ChatGPT specifically for programming tasks? (3) How does prior IT experience moderate the relationship between perceived benefits and adoption intention? We propose AdoptGPT-Prog, an integrated TAM-UTAUT framework incorporating eight predictors\u0026mdash;Perceived Usefulness, Perceived Ease of Use, Trust, Anxiety, Perceived Risk, AI Literacy, Perceived Learning Value, and Hedonic Motivation\u0026mdash;with Gender and IT Experience as moderators. Data from 486 undergraduate programming students (52.3% male, 47.7% female; mean age\u0026thinsp;=\u0026thinsp;21.4 years, SD\u0026thinsp;=\u0026thinsp;2.1) across five Saudi Arabian universities were analyzed using PLS-SEM. The model achieved substantial explanatory power (R\u0026sup2;=0.714), outperforming prior ChatGPT adoption frameworks. Perceived Usefulness emerged as the strongest predictor (β\u0026thinsp;=\u0026thinsp;0.312, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), followed by Trust (β\u0026thinsp;=\u0026thinsp;0.203) and Perceived Ease of Use (β\u0026thinsp;=\u0026thinsp;0.187), while Perceived Risk (β=-0.156) functioned as a significant barrier. Notably, Anxiety showed no direct effect on intention (β\u0026thinsp;=\u0026thinsp;0.012, p\u0026thinsp;=\u0026thinsp;0.774), though gender moderated this relationship, with female students exhibiting stronger anxiety-related inhibition. Multi-group analysis revealed that females showed stronger Perceived Ease of Use effects (β\u0026thinsp;=\u0026thinsp;0.279 vs. β\u0026thinsp;=\u0026thinsp;0.112), while high-IT-experience students weighted Perceived Usefulness more heavily (β\u0026thinsp;=\u0026thinsp;0.412 vs. β\u0026thinsp;=\u0026thinsp;0.234). These findings provide empirically-grounded, actionable recommendations for differentiated instructional strategies when integrating generative AI tools into programming curricula.\u003c/p\u003e","manuscriptTitle":"Understanding Factors Influencing Students Intention to Use ChatGPT for Learning Programming with Gender and IT Experience as Moderators Based on AdoptGPT Prog Conceptual Model","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-16 08:21:12","doi":"10.21203/rs.3.rs-8564643/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-04-14T18:46:39+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"261875674285006724163457852479119118740","date":"2026-03-10T07:31:29+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-06T21:18:02+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"233975169670370820236345693835389051669","date":"2026-03-06T21:14:34+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-08T15:17:15+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"42620397630498382109711484692682682540","date":"2026-02-06T21:48:34+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-02-06T20:30:36+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-16T19:12:36+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-01-15T02:54:17+00:00","index":"","fulltext":""},{"type":"submitted","content":"Discover Education","date":"2026-01-15T02:48:49+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"discover-education","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"diedu","sideBox":"Learn more about [Discover Education](https://www.springer.com/journal/44217)","snPcode":"44217","submissionUrl":"https://submission.nature.com/new-submission/44217/3","title":"Discover Education","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Discover Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"8099044f-4c35-4835-8f86-375d5a65fa51","owner":[],"postedDate":"February 16th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-02-16T08:21:12+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-16 08:21:12","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8564643","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8564643","identity":"rs-8564643","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}