Development and Validation of a Scale Assessing College Students’ Negative Attitudes Toward Generative AI-Assisted Academic Writing

Development and Validation of a Scale Assessing College Students’ Negative Attitudes Toward Generative AI-Assisted Academic Writing | 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 Article Development and Validation of a Scale Assessing College Students’ Negative Attitudes Toward Generative AI-Assisted Academic Writing Yanchao Yang, Zhe Shi, Tianxue Cui, Xinxin Yang, Zehan Tan This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6974623/v1 This work is licensed under a CC BY 4.0 License Status: Under Revision Version 1 posted 10 You are reading this latest preprint version Abstract This study adopted a multi-stage research design to develop and validate a scale measuring university students’ attitudes toward the application of generative artificial intelligence (Generative AI) in academic writing. Based on preliminary interviews and qualitative analysis, three core dimensions were identified: language homogenization, thought outsourcing, and identity ambiguity, reflecting the potential impacts of AI on language style, cognitive processes, and authorship identity, respectively. A total of 2,019 participants were recruited. During scale development, item analysis and exploratory factor analysis (EFA) were first conducted. Parallel analysis confirmed the appropriateness of a three-factor structure. Subsequent confirmatory factor analysis (CFA) demonstrated good model fit, as well as strong convergent and discriminant validity. Finally, multi-group confirmatory factor analysis (MG-CFA) further supported the structural invariance of the scale across subgroups defined by gender, birthplace, and educational level, indicating its broad applicability. Despite its solid reliability, validity and broad applicability, the scale still has limitations, such as the need for longitudinal studies, predictive validity testing, and validation using multiple methods. Overall, the scale enriches the understanding of user attitudes toward technology beyond the traditional Affective–Behavioral–Cognitive model, and offers concrete educational insights for the responsible integration of generative AI in academic writing. Social science/Education Social science/Science technology and society attitude Generative AI-assisted academic writing scale development scale validation Figures Figure 1 Introduction In recent years, generative artificial intelligence technology has undergone rapid development (Reddy et al., 2025 ). Driven by breakthroughs in deep learning and large-scale data training, powerful large-scale pre-trained language models such as GPT and LLaMA have emerged, revolutionizing the field of natural language processing (Lauriola et al., 2022 ; Paass & Giesselbach, 2023 ). These technologies have continuously iterated and upgraded at an astonishing pace, not only leading the forefront of AI research but also triggering transformative waves across multiple industries, including medical field (Ali et al., 2023 ), healthcare (Burns et al., 2024 ; Xu & Wang, 2024 ), finance (Puchakayala, 2024 ), media (Brüns & Meißner, 2024 ; Feng et al., 2025 ), and manufacturing (Ghobakhloo et al., 2024 ), profoundly changing the way people work and live. These transformations across various sectors have not only driven industry efficiency and innovation but also prompted the education field to actively embrace generative artificial intelligence technology (Arinushkina, 2025 ; Bozkurt & Sharma, 2024 ; Saúde et al., 2024 ). As a critical link in talent cultivation and social development, education is facing unprecedented opportunities and challenges brought by artificial intelligence (Bobula, 2024 ; Wu et al., 2023 ). Particularly in the realm of academic writing, generative AI technology, with its groundbreaking advancements, is reshaping the academic landscape (Lorenc-Kukula, 2025 ; Mahdi & Alkhateeb, 2025 ; Mondal, 2025 ). Since the Transformer architecture revolutionized natural language processing, large-scale pre-trained models such as the GPT series and LLaMA have undergone continuous iterations, giving rise to a diverse range of writing assistance tools. For instance, GenAI-assisted writing tools assist students in conducting literature reviews by identifying pertinent scholarly articles (Behrooz et al., 2023 ; van Mossel et al., 2025 ), supplying contextual information related to writing topics (Chichekian & Benteux, 2022 ), summarizing key content (Behrooz et al., 2023 ), and generating personalized suggestions based on individual preferences and search patterns (Chichekian & Benteux, 2022 ). Additionally, an GenAI-driven chatbot can support students in creating outlines for writing assignments (Kim et al., 2025 ). By improving the structure of their writing and creating a more interactive learning environment, the AI fosters effective writing habits. Through targeted questions and prompts, the chatbot encourages critical thinking, offering opportunities for reflection and enhancing self-regulation in the writing process (Lin & Chang, 2020 ). Furthermore, AI tools, such as ChatGPT, support users in brainstorming ideas and receiving feedback on their writing, thereby enhancing the overall writing process (Atlas, 2023 ). In addition, GenAI tools are increasingly recognized as valuable research assistants, capable of generating ideas, synthesizing complex information, and summarizing large volumes of text. These functions help researchers analyze data more efficiently and improve the quality of their academic writing (Berg, 2023 ). The widespread application of generative AI technology in academic writing has not only transformed the way academic writing is conducted but also sparked varied attitudes and perspectives toward this emerging technology. Scholars and researchers exhibit a broad range of attitudes toward the integration of generative artificial intelligence in academic writing (Cohen & Moher, 2025 ; Jin et al., 2025 ; Pu et al., 2024 ; Stojanov, 2025 ). An attitude is a person's comprehensive assessment of an object or topic, encompassing both emotional and evaluative aspects (Ajzen, 2001 ; Ajzen & Fishbein, 2000 ).When analyzing attitudes, the ABC (Affect–Behavior–Cognition) model is one of the most commonly used frameworks. It enables researchers to explore attitudes from three dimensions: affective (e.g., emotions such as trust or curiosity), behavioral (e.g., actual usage or intention to use), and cognitive (e.g., perceived usefulness or ease of use) (Breckler, 1984 ; Jackson et al., 1996 ; Millar & Tesser, 1989 ; Wolff et al., 2011 ). While this model provides a useful overarching framework, it tends to remain at a broad level of analysis and lacks the specificity needed to capture the nuanced impacts of generative AI in academic contexts. Moreover, most existing studies focus heavily on the positive attitude (e.g., Atlas, 2023 ; Behrooz et al., 2023 ; van Mossel et al., 2025 )—while paying limited attention to negative perceptions. This imbalance may lead to an overly optimistic portrayal of generative AI adoption, overlooking critical barriers such as ethical concerns, mistrust, or anxiety, which could significantly influence actual user behavior and long-term integration in academic settings. Furthermore, few studies have employed quantitative methods to rigorously verify or measure these negative attitudes. As a result, current understandings remain largely theoretical. The lack of empirical validation not only hinders the development of nuanced theoretical models but also weakens the foundation for evidence-based policymaking and practical interventions, potentially resulting in misaligned strategies that fail to address users’ real concerns or resistance. To this end, this study aims to develop and validate a negative attitude scale toward GenAI-assisted academic writing among university students. Unlike general attitude measures, this scale is specifically designed for the academic writing context, ensuring high relevance and contextual sensitivity. Moreover, it shifts the research focus toward the often-overlooked negative perceptions, such as distrust, fear of over-reliance, and ethical concerns. By adopting a quantitative approach, the scale enables systematic measurement and comparison, providing robust empirical evidence for understanding the complexity of student responses to GenAI. This contributes not only to theory-building but also to the development of more balanced, responsive, and ethical educational policies and practices. Methods Participants The procedures used in this study adhere to the tenets of the Declaration of Helsinki. Prior to data collection, the ethical approval has been granted on 17/12/2024, by Macau Millennium College, with approval number MMCIRB-2024-002. The written consent form was embedded in the online questionnaire, and participants could proceed only after selecting the "Willing to participate" option. To protect participants' privacy, no identifiable information, such as names, contact details, or student ID, was collected. A total of 2,140 individuals accessed the survey, among whom 2,019 consented to participate and completed the questionnaire from March 26 to March 29, 2025. Respondents were from various regions, including China, such as Henan, Hebei, Beijing, Shanghai, Gansu, Guangxi, Heilongjiang, Jilin, Shaanxi, Chongqing, and Australia. The survey was distributed via the Wenjuanxing platform. The estimated completion time was approximately 3 minutes. A multi-stage design was adopted for instrument development and validation. The dataset was randomly divided into three groups. The first group (n = 673) was used for item analysis and exploratory factor analysis. The second group (n = 673) was used to conduct confirmatory factor analysis. The third group was (n = 673) used for multi-group confirmatory factor analysis to examine measurement invariance across different subgroups. A detailed summary of the participants’ demographic characteristics is provided in Table 1 . Table 1 demographic information Demographic Variable Group EFA Dataset CFA Dataset MG- CFA Dataset N % N % N % Gender Male 108 16.00% 141 21.00% 160 23.80% Female 565 84.00% 532 79.00% 513 76.20% Education Undergraduate 639 94.90% 641 95.20% 641 95.20% Graduate 34 5.10% 32 4.80% 32 4.80% Birthplace Urban 201 29.90% 196 29.10% 189 28.10% Rural 472 70.10% 477 70.90% 484 71.90% Analytical procedure A multi-stage design was employed for instrument development and validation. The dataset was randomly partitioned into three subsets to support sequential psychometric analyses and ensure the robustness and generalizability of the findings. In the first stage, item analysis and exploratory factor analysis (EFA) were conducted using the first dataset. Item analysis included high-low group comparison, item-total correlation, and Cronbach’s alpha if item deleted, in order to identify items with low discrimination or internal inconsistency. Items that failed to show adequate discrimination between high and low scoring groups (i.e., with non-significant or weak differences) were considered to have insufficient discriminative power and were removed. Additionally, items with corrected item-total correlations below .30 or whose deletion improved the overall internal consistency were flagged for exclusion. Prior to EFA, the Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s test of sphericity were used to confirm the data’s suitability for factor analysis (Kaiser, 1974 ). EFA was then conducted and the number of factors was determined by parallel analysis with oblique rotation (promax). Parallel analysis was chosen because it compares the observed eigenvalues with those generated from random datasets to reduce the risk of factor over-extraction and improve the accuracy of factor retention decisions. Items with factor loadings below .40, cross-loadings above .30, or weak conceptual relevance were excluded to optimize the factor structure. In the second stage, confirmatory factor analysis (CFA) was conducted using the second dataset to validate the factor structure identified in EFA. Model fit was assessed using a combination of indices (Hu & Bentler, 1999 ; McDonald & Ho, 2002 ): Comparative Fit Index (CFI) and Tucker–Lewis Index (TLI) values ≥ 0.90 indicated acceptable fit, and ≥ 0.95 indicated good fit. Root Mean Square Error of Approximation (RMSEA) values ≤ 0.08 were considered acceptable (≤ 0.06 for good fit), and Standardized Root Mean Square Residual (SRMR) values ≤ 0.05 were also indicative of good model fit. Convergent validity was assessed by examining Average Variance Extracted (AVE) and Composite Reliability (CR); AVE ≥ 0.50 and CR ≥ 0.70 were considered acceptable (Hair et al., 2010 ). Discriminant validity was examined using the Fornell-Larcker criterion, ensuring that the square root of AVE for each construct exceeded the inter-construct correlations (Fornell & Larcker, 1981 ). In the third stage, multi-group confirmatory factor analysis (MG-CFA) was conducted using the third dataset to test measurement invariance across subgroups (e.g., gender, birthplace and educational background). A sequential procedure was employed to test the measurement invariance, including configural, metric, scalar, and strict invariance. Each step imposed increasing constraints on the model parameters. Changes in model fit between nested models were evaluated using ΔCFI, ΔTLI and ΔRMSEA, with ΔCFI and ΔTLI ≤ 0.01 and ΔRMSEA ≤ 0.015 indicating that invariance held at each level (Byrne et al., 1989 ; Chen, 2007 ). This process ensured that the instrument measured the constructs equivalently across different groups. Results Study 1 Item development To develop the initial pool of scale items, we conducted semi-structured interviews with five graduate students who had hands-on experience using generative AI tools to support academic writing. To ensure ethical standards and protect participant privacy, all interviewees were informed that the data collected would be used exclusively for academic purposes, with no personally identifiable information disclosed. The interview protocol focused on participants’ attitudes toward AI-assisted writing, including their motivations for use, perceived benefits and challenges, shifts in writing practices, and views on authorship and identity. Each interview lasted approximately 40 minutes and was audio-recorded and transcribed. The transcripts were analyzed through a multi-stage coding process to ensure analytical rigor. During open coding, the researchers conducted line-by-line reading to extract initial concepts and assign descriptive labels to meaningful text segments. In axial coding, similar codes were grouped and refined to identify underlying patterns, while irrelevant or redundant content was eliminated. Selective coding was then used to synthesize these categories into three overarching themes—language homogenization, thought outsourcing, and identity ambiguity—that reflected participants’ nuanced reflections on the implications of GenAI in academic writing and informed the item development. To enhance interpretive validity, the themes and key interpretations were sent back to the interviewees for member checking, allowing them to confirm accuracy or raise objections where necessary. Content validity Subsequently, to assess content validity, eight experts in education and information technology were invited to evaluate the relevance, clarity, and representativeness of each item on a 4-point Likert scale. The item-level content validity index (I-CVI) was calculated as the proportion of experts rating the item as either 3 or 4. Additionally, the scale-level content validity index (S-CVI) was computed as the average of all I-CVIs. As all items achieved an I-CVI of 0.80 or above—commonly accepted as the minimum threshold for adequate content validity—and the S-CVI exceeded 0.90, no revisions were deemed necessary at this stage. Study 2 Item analysis Independent samples t -tests were performed between the upper 27% group (total scores ≥ 64) and the lower 27% group (total scores ≤ 48), revealing that all items differed significantly at the 95% confidence interval, thereby demonstrating strong discriminative validity. Furthermore, item-total correlations, measured by Pearson’s correlation coefficients, ranged from 0.742 to 0.901 ( p = 0.01), indicating robust associations between individual items and the overall scale score. The internal consistency reliability, assessed by Cronbach’s alpha, was 0.975. Deletion of any single item did not result in a significant increase in the alpha coefficient, confirming the high reliability and homogeneity of the instrument. Exploratory factor analysis The suitability of the data for exploratory factor analysis (EFA) was first evaluated. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy was exceptionally high at 0.965, indicating excellent factorability of the correlation matrix. Additionally, Bartlett’s test of sphericity was significant, χ²(120) = 15,019.689, p < 0.001, rejecting the null hypothesis that the correlation matrix is an identity matrix and confirming the appropriateness of conducting factor analysis. Parallel analysis was then performed to determine the optimal number of factors to retain. Subsequently, an exploratory factor analysis was conducted employing an oblique Promax rotation, which allows for correlation among factors. Parallel analysis is preferred because it provides a more accurate and objective criterion by comparing the observed eigenvalues with those obtained from randomly generated data, thereby reducing the risk of over-extraction. During the EFA process, items were considered for deletion based on the following criteria: low factor loadings (below 0.40), significant cross-loadings (loadings ≥ 0.30 on multiple factors), and lack of theoretical consistency with the intended construct. However, none of the items met these criteria for removal, and therefore, all original items were retained in the final factor structure. The analysis yielded a three-factor solution, which aligned well with the theoretically hypothesized factor structure. The extracted factors collectively explained a substantial proportion of the total variance as shown in Table 2 , and factor loadings demonstrated clear and interpretable item-factor associations consistent with the predefined dimensions. Table 2 Exploratory factor analysis results Item LH TO IA Communality Cronbach’s alpha coefficient Variance explained LH1 0.912 0.818 0.937 34% LH2 0.881 0.786 LH3 0.864 0.832 LH4 0.768 0.731 TO1 0.881 0.826 0.969 29.90% TO2 0.884 0.873 TO3 0.755 0.786 TO4 0.796 0.847 TO5 0.807 0.86 TO6 0.756 0.856 IA1 0.931 0.878 0.977 20.30% IA2 0.901 0.884 IA3 0.872 0.883 IA4 0.912 0.89 IA5 0.839 0.863 IA6 0.856 0.874 Note: IH = language homogenization, TO = thought outsourcing and IA = identity ambiguity; Factor loadings below 0.40 are not displayed. Study 3: Confirmatory factor analysis To test whether the observed variables adequately reflected the hypothesized latent constructs, confirmatory factor analysis (CFA) was conducted on the second dataset using the CB-SEM module in SmartPLS 4.0. The results in Table 3 indicated a good model fit. Table 3 Model fit statistics χ2 df p χ2/df RMSEA 90% CI GFI SRMR TLI CFI 382.217 101 < 0.01 3.784 0.064[0.058–0.071] 0.932 0.016 0.978 0.981 Note: RMSEA = Root Mean Square Error of Approximation 90% CI = 90% Confidence Interval (for RMSEA) GFI = Goodness-of-Fit Index SRMR = Standardized Root Mean Square Residual TLI = Tucker-Lewis Index (also known as Non-Normed Fit Index, NNFI) CFI = Comparative Fit Index Convergent validity To assess convergent validity, composite reliability (CR), average variance extracted (AVE), and standardized factor loadings were examined. As shown in Table 4 , all constructs demonstrated strong internal consistency, with CR values ranging from 0.938 to 0.975, exceeding the recommended threshold of 0.70. The AVE values ranged from 0.791 to 0.867, all above the minimum acceptable value of 0.50, indicating that each construct explains a substantial proportion of variance in its observed indicators. In addition, all standardized factor loadings were statistically significant and ranged from 0.875 to 0.945 (also see Fig. 1 ), further supporting the convergent validity of the scale. Table 4 CR and AVE results Factor Composite reliability Average variance extracted identity ambiguity 0.975 0.867 language homogenization 0.938 0.791 thought outsourcing 0.973 0.855 Discriminant validity Discriminant validity was assessed using the Fornell-Larcker criterion. As shown in Table 5 , the square root of the AVE for each construct (shown on the diagonal) exceeded its correlations with the other constructs, indicating adequate discriminant validity. Table 5 Fornell-Larcker results Factor 1 2 3 1. identity ambiguity 0.931 2. language homogenization 0.689 0.890 3. thought outsourcing 0.887 0.699 0.925 Study 4 Multi-group Confirmatory Factor Analysis The multi-group confirmatory factor analysis (MG-CFA) was conducted using the third dataset to assess measurement invariance across birthplace, sex, and education level. Sequential tests of configural, metric, scalar, and residual invariance all demonstrated good model fit, with no significant declines when equality constraints were imposed. Fit indices as shown in Table 6 remained within acceptable thresholds (e.g., ΔCFI < 0.01, ΔTFI < 0.01, ΔRMSEA < 0.015), supporting full measurement invariance across these demographic groups. These findings indicate that the measurement model operates equivalently across subgroups defined by birthplace, sex, and education level, allowing for meaningful group comparisons. Table 6 MG-CFA results Demographic Variable Model χ2 df χ2/df Δχ2 Δ df p CFI ΔCFI TLI ΔTLI RMSEA ΔRMSEA Birthplace Unconstrained 658.548 202.000 3.260 — — — 0.970 — 0.964 — 0.058 — Configural invariance 671.728 215.000 3.124 13.180 13.000 0.434 0.970 0.000 0.966 0.002 0.056 -0.002 Metric invariance 688.453 231.000 2.980 16.725 16.000 0.404 0.970 0.000 0.969 0.003 0.054 -0.002 Scalar invariance 694.789 237.000 2.932 6.336 6.000 0.387 0.970 0.000 0.969 0.000 0.054 0.000 Residual invariance 755.253 253.000 2.985 60.464 16.000 < 0.001 0.967 -0.003 0.969 0.000 0.054 0.000 Sex Unconstrained 600.549 202 2.973 — — — 0.973 — 0.968 — 0.054 — Configural invariance 608.032 215 2.828 7.483 13 0.876 0.974 0.001 0.971 0.003 0.052 -0.002 Metric invariance 661.399 231 2.863 53.367 16 < 0.001 0.971 -0.003 0.97 -0.001 0.053 0.001 Scalar invariance 695.124 237 2.933 33.725 6 < 0.001 0.969 -0.002 0.969 -0.001 0.054 0.001 Residual invariance 787.328 253 3.112 92.204 16 < 0.001 0.964 -0.005 0.966 -0.003 0.056 0.002 Education Unconstrained 580.597 202.000 2.874 — — — 0.975 — 0.970 — 0.053 — Configural invariance 591.988 215.000 2.753 11.391 13.000 0.578 0.975 0.000 0.972 0.002 0.051 -0.002 Metric invariance 603.929 231.000 2.614 11.941 16.000 0.748 0.975 0.000 0.974 0.002 0.049 -0.002 Scalar invariance 606.879 237.000 2.561 2.950 6.000 0.815 0.976 0.001 0.975 0.001 0.048 -0.001 Residual invariance 737.815 253.000 2.916 130.936 16.000 < 0.001 0.968 -0.008 0.970 -0.005 0.053 0.005 Discussion In this study, we adopted a multi-study design to develop and validate a scale measuring university students’ attitudes toward the application of generative artificial intelligence (Generative AI) in academic writing. A total of 2,019 participants were recruited, providing a large and representative sample. Based on preliminary interviews and qualitative analysis, three core dimensions were identified through coding and thematic extraction: language homogenization (4 items), thought outsourcing (6 items), and identity ambiguity (6 items). These dimensions comprehensively capture students’ attitudes toward the potential impacts of generative AI on academic writing, from the perspectives of language style, cognitive processes, and self-identity. Subsequently, item analysis and exploratory factor analysis (EFA) were conducted. Parallel analysis confirmed the appropriateness of a three-factor structure, which aligned closely with the theoretically hypothesized dimensions. All items showed strong loadings and clear structural distinctions within their respective factors, with no items needing to be removed. Further confirmatory factor analysis (CFA) provided robust support for the structural model, demonstrating good model fit as well as significant convergent and discriminant validity, thereby confirming the construct validity of the scale. In addition, multi-group confirmatory factor analysis (MG-CFA) indicated that the scale exhibited structural invariance across subgroups defined by birthplace, gender, and educational level, confirming its suitability for meaningful cross-group comparisons and its broad applicability across diverse demographic profiles. The language homogenization factor captures students’ perceptions of how generative AI contributes to the stylistic standardization and emotional flattening of academic writing. The items under this dimension reflect a clear pattern of linguistic convergence, including the frequent adoption of AI-recommended passive constructions, complex syntactic patterns, and formalized paragraph structures. Additionally, participants expressed a tendency to use high-frequency, generic vocabulary suggested by AI, which diminishes their ability to adapt language to diverse topics or express personal voice. Importantly, the avoidance of subjective tone and emotional expression—as noted in one item—underscores a shift toward neutral, impersonal academic discourse, making writing appear emotionally flat or detached. Together, these items consistently reflect a core concern: that excessive reliance on generative AI may gradually erode the individuality, creativity, and affective richness of academic writing, leading to stylistic homogenization across authors and contexts. The thought outsourcing dimension is similarly well represented, with items reflecting a progressive delegation of core cognitive processes—such as ideation, critical reasoning, and coherence checking—to generative AI. This factor captures not only behavioral reliance (e.g., default use of AI-generated arguments or structure) but also a deeper psychological disengagement from academic authorship, including diminished agency, reduced creativity, and a sense of detachment from one’s own thought process. Lastly, items under the identity ambiguity factor reflect students’ growing uncertainty and internal conflict regarding their academic roles and authorship legitimacy. These items articulate the blurred boundaries between human and AI contributions in academic writing, including concerns about intellectual ownership, ethical detachment, and the performative nature of authorship in an AI-mediated academic environment. The metaphorical expressions used by participants—such as “ghostwriter,” “academic tailor,” or “Turing Test” participant—not only highlight the depth of this identity confusion but also resonate strongly with contemporary discourse on authorship and AI ethics. Implications This study advances the theoretical understanding of attitudes toward technology by moving beyond the generalist Affect–Behavior–Cognition (ABC) model. Rather than focusing solely on affective reactions or behavioral intentions, we propose a domain-specific framework consisting of language homogenization, thought outsourcing, and identity ambiguity—three dimensions that capture the cognitive, linguistic, and identity-related dilemmas students face when engaging with generative AI in academic writing. This situated approach provides a more nuanced and explanatory account of user attitudes, highlighting how AI reshapes students’ expressive style, cognitive autonomy, and sense of authorship. Compared to general models that assess broad sentiments like willingness to use or perceived usefulness, our framework offers deeper insight into the psychological and behavioral transformations emerging in AI-integrated academic contexts. From an educational perspective, the scale provides actionable insights for educators and institutions seeking to responsibly integrate generative AI into academic settings. By identifying specific challenges—such as linguistic standardization, diminished critical engagement, and blurred authorship—educators can better anticipate the unintended consequences of AI-assisted writing. These findings underscore the need for pedagogical strategies that go beyond teaching students how to use AI tools, and instead focus on when, why, and to what extent they should be used. Curriculum designers might incorporate critical AI literacy, reflective writing tasks, and authorship ethics into writing instruction to help students retain cognitive ownership and identity coherence. The scale also serves as a diagnostic tool for monitoring students’ evolving attitudes and guiding adaptive teaching interventions in rapidly changing technological landscapes. Limitations and suggestions Although this study employed rigorous psychometric procedures—including exploratory and confirmatory factor analyses—to establish the structural validity of the scale, future research could benefit from methodological triangulation. For instance, applying item response theory (IRT) could provide more detailed insights into individual item functioning across different subgroups, while network analysis might reveal latent inter-item relationships and centrality patterns. These complementary approaches would offer a more nuanced understanding of the scale’s dimensional structure and enhance the robustness of its construct validity. The current study is cross-sectional in nature, which constrains our ability to examine the developmental trajectory of students’ attitudes toward generative AI in academic writing. Given that attitudes may evolve as students gain more exposure, experience, or institutional guidance regarding AI technologies, future research should adopt longitudinal designs. Tracking attitude changes over time would help determine the stability or transformation of the identified dimensions (e.g., whether thought outsourcing intensifies or diminishes with AI familiarity), and provide a dynamic view of student adaptation in AI-mediated learning environments. While the scale effectively captures attitudinal dimensions, its predictive validity remains to be established. Specifically, it is unclear how these attitudes—such as concerns over identity ambiguity or tendencies toward thought outsourcing—translate into actual academic writing behaviors or outputs. Future studies should examine the extent to which these factors predict writing performance indicators such as originality, coherence, ethical conduct, or engagement with feedback. Such investigations would not only validate the scale’s practical utility but also offer educators diagnostic insights for intervention and support. Conclusion This study developed and validated a scale measuring university students’ attitudes toward the use of generative AI in academic writing, encompassing three key dimensions: language homogenization, thought outsourcing, and identity ambiguity. The scale demonstrated strong reliability and validity across diverse groups. These findings contribute to a deeper theoretical understanding and offer practical guidance for educators in designing responsible and reflective AI-integrated writing instruction. Declarations Ethical approval The procedures used in this study adhere to the tenets of the Declaration of Helsinki. Prior to data collection, the ethical approval has been granted by Macau Millennium College, with approval number MMCIRB-2024-002. Informed consent The written consent form was embedded in the online questionnaire, and participants could proceed only after selecting the "Willing to participate" option. To protect participants' privacy, no identifiable information, such as names, contact details, or student ID, was collected. Participants were assured of anonymity and confidentiality, with the understanding that their responses would be used solely in aggregated form for academic publications and presentations. They were also informed that they could withdraw from the study at any time, and in such cases, their data would be securely destroyed upon request. Competing interests The authors declare no competing interests. Funding Declaration This study received no funding. Author Contribution Author ContributionsConceptualization: YYC and SZMethodology: YYC and CTXSoftware: YYC and CTXValidation: YYC and TZHFormal analysis: YYC and SZData curation: YYC and SZWriting – original draft preparation: YYCWriting – review and editing: YXX and TZHVisualization: YXXSupervision: YYCProject Administration: YYC Data Availability Data is provided within the manuscript or supplementary information files References Ajzen I (2001) Nature and Operation of Attitudes. Ann Rev Psychol 52(1):27–58. https://doi.org/10.1146/annurev.psych.52.1.27 Ajzen I, Fishbein M (2000) Attitudes and the Attitude-Behavior Relation: Reasoned and Automatic Processes. Eur Rev Social Psychol 11(1):1–33. https://doi.org/10.1080/14792779943000116 Ali H, Rehmani MH, Shāh SZ (2023) & service), S. (Online. Advances in Deep Generative Models for Medical Artificial Intelligence (1st ed. 20). Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-46341-9 Arinushkina AA (2025) Integration strategies of generative AI in higher education. Integration strategies of generative artificial intelligence in higher education. IGI Global Atlas S (2023) ChatGPT for higher education and professional development: A guide to conversational AI . https://digitalcommons.uri.edu/cba_facpubs/548 Behrooz H, Lipizzi C, Korfiatis G, Ilbeigi M, Powell M, Nouri M (2023) Towards Automating the Identification of Sustainable Projects Seeking Financial Support: An AI-Powered Approach. Sustainability 15(12):9701. https://doi.org/10.3390/su15129701 Berg C (2023) The case for generative AI in scholarly practice . https://doi.org/10.2139/ssrn.4407587 Bobula M (2024) Generative artificial intelligence (AI) in higher education: a comprehensive review of challenges, opportunities, and implications. Journal of Learning Development in Higher Education , 30 . https://doi.org/10.47408/jldhe.vi30.1137 Bozkurt A, Sharma RC (2024) Transforming education with generative AI. In Transforming education with generative artificial intelligence . IGI Global, Engineering Science Reference. https://doi.org/10.4018/979-8-3693-1351-0 Breckler SJ (1984) Empirical validation of affect, behavior, and cognition as distinct components of attitude. J Personal Soc Psychol 47(6):1191 Brüns JD, Meißner M (2024) Do you create your content yourself? Using generative artificial intelligence for social media content creation diminishes perceived brand authenticity. J Retailing Consumer Serv 79:103790. https://doi.org/10.1016/j.jretconser.2024.103790 Burns B, Nemelka B, Arora A (2024) Practical implementation of generative artificial intelligence systems in healthcare: A United States perspective Practical implementation of generative artificial intelligence systems in healthcare: A United States perspective. Future Healthc J 11(3):100166 Byrne BM, Shavelson RJ, Muthén B (1989) Testing for the equivalence of factor covariance and mean structures: The issue of partial measurement invariance. Psychol Bull 105(3):456–466. https://doi.org/10.1037/0033-2909.105.3.456 Chen FF (2007) Sensitivity of goodness of fit indexes to lack of measurement invariance. Struct Equation Modeling: Multidisciplinary J 14(3):464–504 Chichekian T, Benteux B (2022) The potential of learning with (and not from) artificial intelligence in education. Front Artif Intell 5:903051. https://doi.org/10.3389/frai.2022.903051 Cohen JF, Moher D (2025) Generative artificial intelligence and academic writing: friend or foe? J Clin Epidemiol 179:111646. https://doi.org/10.1016/j.jclinepi.2024.111646 Feng W, Li Y, Ma C, Yu L (2025) From ChatGPT to Sora: Analyzing Public Opinions and Attitudes on Generative Artificial Intelligence in Social Media. IEEE Access 13:14485–14498. https://doi.org/10.1109/ACCESS.2025.3530683 Fornell C, Larcker DF (1981) Evaluating structural equation models with unobservable variables and measurement error. J Mark Res 18(1):39–50. https://doi.org/10.2307/3151312 Ghobakhloo M, Fathi M, Iranmanesh M, Vilkas M, Grybauskas A, Amran A (2024) Generative artificial intelligence in manufacturing: opportunities for actualizing Industry 5.0 sustainability goals. J Manuf Technol Manage 35(9):94–121. https://doi.org/10.1108/JMTM-12-2023-0530 Hair J, Black W, Babin B, Anderson R (2010) Multivariate data analysis (7th Editio). Prentice-Hall, Inc Hu LT, Bentler PM (1999) Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria Versus New Alternatives. Struct Equation Modeling: Multidisciplinary J 6(1):1–55. https://doi.org/10.1080/10705519909540118 Jackson LA, Hodge CN, Gerard DA, Ingram JM, Ervin KS, Sheppard LA (1996) Cognition, Affect, and Behavior in the Prediction of Group Attitudes. Personal Soc Psychol Bull 22(3):306–316. https://doi.org/10.1177/0146167296223009 Jin F, Sun L, Pan Y, Lin C-H (2025) High heels, compass, spider-man, or drug? Metaphor analysis of generative artificial intelligence in academic writing. Computers Educ 228:105248. https://doi.org/10.1016/j.compedu.2025.105248 Kaiser HF (1974) An index of factorial simplicity. Psychometrika 39(1):31–36. https://doi.org/10.1007/BF02291575 Kim J, Yu S, Detrick R, Li N (2025) Exploring students’ perspectives on Generative AI-assisted academic writing. Educ Inform Technol 30(1):1265–1300. https://doi.org/10.1007/s10639-024-12878-7 Lauriola I, Lavelli A, Aiolli F (2022) An introduction to Deep Learning in Natural Language Processing: Models, techniques, and tools. Neurocomputing (Amsterdam) 470:443–456. https://doi.org/10.1016/j.neucom.2021.05.103 Lin MP-C, Chang D (2020) Enhancing Post secondary Writers’ Writing Skills with a Chatbot: A Mixed-Method Classroom Study. Educational Technol ༆ Soc 23(1):78–92. https://doi.org/10.30191/ETS.202001_23(1).0006 Lorenc-Kukula K (2025) Cutting-edge AI tools revolutionizing scientific research in life sciences. Biotechnologia (Poznan) 106(1):77–102. https://doi.org/10.5114/bta/200803 Mahdi HS, Alkhateeb A (2025) Revolutionising Essay Evaluation: A Cutting-Edge Rubric for AI-Assisted Writing. Int J Computer-Assisted Lang Learn Teach 15(1):1–19. https://doi.org/10.4018/IJCALLT.368226 McDonald RP, Ho M-HR (2002) Principles and practice in reporting structural equation analyses. Psychol Methods 7(1):64–82. https://doi.org/10.1037/1082-989X.7.1.64 Millar MG, Tesser A (1989) The effects of affective-cognitive consistency and thought on the attitude-behavior relation. J Exp Soc Psychol 25(2):189–202. https://doi.org/10.1016/0022-1031(89)90012-7 Mondal H (2025) The Future of Writing: How Artificial Intelligence is Shaping the Way We Write. Indian J Vascular Endovascular Surg 12(1):74–75. https://doi.org/10.4103/ijves.ijves_160_24 Paass G, Giesselbach S (2023) Foundation Models for Natural Language Processing: Pre-trained Language Models Integrating Media (1st ed. 20). Springer International Publishing. https://doi.org/10.1007/978-3-031-23190-2 Pu Z, Shi C, Jeon CO, Fu J, Liu S, Lan C, Yao Y, Liu Y, Jia B (2024) ChatGPT and generative AI are revolutionizing the scientific community: A Janus-faced conundrum. IMeta 3(2):e178–n. /a https://doi.org/10.1002/imt2.178 Puchakayala PRA (2024) Generative Artificial intelligence Applications in Banking and Finance sector. World J Adv Res Reviews 23(1):3105–3120. https://doi.org/10.30574/wjarr.2024.23.1.1999 Reddy P, Chand K, Sharma K, Sharma B, Sharma S (2025) Evolution of Generative Artificial Intelligence: A Review of the Developed and Developing. Eng Sci. https://doi.org/10.30919/es1529 Saúde S, Barros JP, Almeida I (2024) Impacts of Generative Artificial Intelligence in Higher Education: Research Trends and Students’ Perceptions. Social Sci (Basel) 13(8):410. https://doi.org/10.3390/socsci13080410 Stojanov A (ed) (2025) Editors’ Newsroom: Polishing our words or replacing our voices? Generative artificial intelligence in academic writing. Social Behavior and Personality , 53 (3), 1–2. https://doi.org/10.2224/sbp14842 van Mossel S, Oude-Wolcherink MJ, de Feria Cardet RE, de Geus-Oei L-F, Vriens D, Koffijberg H, Saing S (2025) Artificial Intelligence as a New Research Ally? Performing AI-Assisted Systematic Literature Reviews in Health Economics: AI as an Ally? Performing AI-Assisted Systematic Literature Reviews in Health Economics. PharmacoEconomics 43(6):647–650. https://doi.org/10.1007/s40273-025-01481-4 Wolff K, Nordin K, Brun W, Berglund G, Kvale G (2011) Affective and cognitive attitudes, uncertainty avoidance and intention to obtain genetic testing: An extension of the Theory of Planned Behaviour. Psychol Health 26(9):1143–1155. https://doi.org/10.1080/08870441003763253 Wu D, Chen L, Han Y (2023) Challenges that generative artificial intelligence poses to higher education and management and the countermeasures— A visualized analysis of literature from 2019 to 2023 using CiteSpace. 733–738. https://doi.org/10.1145/3660043.3660174 Xu R, Wang Z (2024) Generative artificial intelligence in healthcare from the perspective of digital media: Applications, opportunities and challenges. Heliyon, 10(12), e32364 Additional Declarations No competing interests reported. Supplementary Files EFAdataset.xlsx CFAdataset.xlsx MGCFAdataset.xlsx Appendix.docx Cite Share Download PDF Status: Under Revision Version 1 posted Editorial decision: Revision requested 06 Jan, 2026 Reviews received at journal 26 Nov, 2025 Reviews received at journal 03 Nov, 2025 Reviewers agreed at journal 02 Nov, 2025 Reviewers agreed at journal 02 Nov, 2025 Reviewers invited by journal 18 Sep, 2025 Editor assigned by journal 18 Sep, 2025 Editor invited by journal 18 Sep, 2025 Submission checks completed at journal 04 Sep, 2025 First submitted to journal 04 Sep, 2025 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-6974623","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":521707807,"identity":"87b88c99-b61f-480b-be55-cfc2deba681c","order_by":0,"name":"Yanchao Yang","email":"","orcid":"","institution":"Macau Millennium College","correspondingAuthor":false,"prefix":"","firstName":"Yanchao","middleName":"","lastName":"Yang","suffix":""},{"id":521707810,"identity":"1d05f483-623b-4370-9339-3ac7c6972ebd","order_by":1,"name":"Zhe Shi","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyUlEQVRIiWNgGAWjYBACfvbm479/VEjwsLE3EKlFsudYgjTDGRsZfp4DRGoxuJGjIM3YlmYjOSOBWFsachiMC9gO8xjcfLzxBkONTTRBLfwMZw8kz+ABarmdVmzBcCwtt4GgLY19CQd4JEBacswkGBsOE9ZicJjHsIHHAOSwM8RqOcZjzMyTkMYjOYOHSC2SPWxpjDMO2PDw8wD9kkCMX/jlHx9j+PhPwp6N/fDGGx9qbAhrQXGkRAIpyiFaSNUxCkbBKBgFIwMAAN7YPtuC3QOUAAAAAElFTkSuQmCC","orcid":"","institution":"Hebei Medical University","correspondingAuthor":true,"prefix":"","firstName":"Zhe","middleName":"","lastName":"Shi","suffix":""},{"id":521707811,"identity":"7902451e-d09d-419f-8a92-4c65fbde518d","order_by":2,"name":"Tianxue Cui","email":"","orcid":"","institution":"Dalian Maritime University","correspondingAuthor":false,"prefix":"","firstName":"Tianxue","middleName":"","lastName":"Cui","suffix":""},{"id":521707812,"identity":"1477dae6-4719-4e09-86d8-2f43d019b1cd","order_by":3,"name":"Xinxin Yang","email":"","orcid":"","institution":"Macau Millennium College","correspondingAuthor":false,"prefix":"","firstName":"Xinxin","middleName":"","lastName":"Yang","suffix":""},{"id":521707813,"identity":"b404f951-e15d-4819-afdc-57db2c43595c","order_by":4,"name":"Zehan Tan","email":"","orcid":"","institution":"Macau Millennium College","correspondingAuthor":false,"prefix":"","firstName":"Zehan","middleName":"","lastName":"Tan","suffix":""}],"badges":[],"createdAt":"2025-06-25 12:23:16","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6974623/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6974623/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92477937,"identity":"9ac42b82-21d8-4640-9d52-5c8d7fa39e0c","added_by":"auto","created_at":"2025-09-30 07:25:40","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":128822,"visible":true,"origin":"","legend":"","description":"","filename":"anonymisedmanuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/aad1cddc389c25dd1e5368c5.docx"},{"id":92475621,"identity":"93b84b11-dab6-4771-a3ab-bf0c147300dc","added_by":"auto","created_at":"2025-09-30 07:17:40","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6632,"visible":true,"origin":"","legend":"","description":"","filename":"3e64852f4fb1432bbf71c43ac332ef4e.json","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/8c2a6d269954cb8e6714dbe8.json"},{"id":92475622,"identity":"add18224-f1b5-4fd2-9cda-bac31104b1f4","added_by":"auto","created_at":"2025-09-30 07:17:40","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":19069,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/8b79d29de043f24ba214588c.docx"},{"id":92475635,"identity":"a9fd5127-1e9d-4004-a0e8-7bd629ebe31d","added_by":"auto","created_at":"2025-09-30 07:17:43","extension":"xlsx","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":65618,"visible":true,"origin":"","legend":"","description":"","filename":"CFAdataset.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/bab0a72495d6ef28599d33f4.xlsx"},{"id":92475634,"identity":"ff39d908-8ad1-4238-995e-e399c1e0e64f","added_by":"auto","created_at":"2025-09-30 07:17:41","extension":"xlsx","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":67238,"visible":true,"origin":"","legend":"","description":"","filename":"EFAdataset.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/45b06921b16b9ef646a540da.xlsx"},{"id":92477939,"identity":"15063198-d2de-45e4-802d-d076c09959c4","added_by":"auto","created_at":"2025-09-30 07:25:40","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":69294,"visible":true,"origin":"","legend":"","description":"","filename":"MGCFAdataset.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/038e21e1663161360c856335.xlsx"},{"id":92475632,"identity":"2acbfa76-ffeb-4a72-a5df-e60461af1a0d","added_by":"auto","created_at":"2025-09-30 07:17:40","extension":"xml","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":130251,"visible":true,"origin":"","legend":"","description":"","filename":"3e64852f4fb1432bbf71c43ac332ef4e1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/2e4b4df13d8329e2d90f7f6f.xml"},{"id":92477938,"identity":"7bc59d2d-308e-4e91-967c-3d798b7be0ac","added_by":"auto","created_at":"2025-09-30 07:25:40","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":43957,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/425b22c5175904761a30fea0.png"},{"id":92475629,"identity":"fd1a41a3-f884-4b35-9b10-b2c94ad9e097","added_by":"auto","created_at":"2025-09-30 07:17:40","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12096,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/ad9a31296b928607a8dac151.png"},{"id":92475630,"identity":"d69134b6-dfe0-4161-b095-3be853fa104e","added_by":"auto","created_at":"2025-09-30 07:17:40","extension":"xml","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":128583,"visible":true,"origin":"","legend":"","description":"","filename":"3e64852f4fb1432bbf71c43ac332ef4e1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/bd42bdfb17388f88aee86619.xml"},{"id":92475631,"identity":"537bfbd4-a948-49b6-899e-2c074e268fbe","added_by":"auto","created_at":"2025-09-30 07:17:40","extension":"html","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":133726,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/0bb15bee823ed0ba256da0bc.html"},{"id":92475619,"identity":"ae110685-7eaf-47cc-bd40-9b10ec5c2cfc","added_by":"auto","created_at":"2025-09-30 07:17:40","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":53096,"visible":true,"origin":"","legend":"\u003cp\u003eConfirmatory factor analysis of the three-factor model\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/095c290f51381f936e922753.png"},{"id":92478572,"identity":"94bc81ab-18a4-4b3d-991d-4fd3cdabfa0c","added_by":"auto","created_at":"2025-09-30 07:33:40","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":920836,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/67eab7b8-9e35-4939-9748-9bdb385225e3.pdf"},{"id":92477936,"identity":"3f30e847-4c1f-4e80-b4d4-9e056f6f9cad","added_by":"auto","created_at":"2025-09-30 07:25:40","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":67238,"visible":true,"origin":"","legend":"","description":"","filename":"EFAdataset.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/e758385a6bcab2729bdc2357.xlsx"},{"id":92475625,"identity":"59ad3943-affa-47bd-9187-98db950ae22c","added_by":"auto","created_at":"2025-09-30 07:17:40","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":65618,"visible":true,"origin":"","legend":"","description":"","filename":"CFAdataset.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/1f5dd423ce05c6c15dbb8b2e.xlsx"},{"id":92475624,"identity":"411a007d-f8eb-42a6-851d-1ed19b48eb51","added_by":"auto","created_at":"2025-09-30 07:17:40","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":69294,"visible":true,"origin":"","legend":"","description":"","filename":"MGCFAdataset.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/ca8532fe2a8eff93f1ed4c5a.xlsx"},{"id":92475628,"identity":"27b7cf4e-b48d-480f-9095-147c0fe48c38","added_by":"auto","created_at":"2025-09-30 07:17:40","extension":"docx","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":19069,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-6974623/v1/024fabf30cea3d3d7738a4ed.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Development and Validation of a Scale Assessing College Students’ Negative Attitudes Toward Generative AI-Assisted Academic Writing","fulltext":[{"header":"Introduction","content":"\u003cp\u003eIn recent years, generative artificial intelligence technology has undergone rapid development (Reddy et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Driven by breakthroughs in deep learning and large-scale data training, powerful large-scale pre-trained language models such as GPT and LLaMA have emerged, revolutionizing the field of natural language processing (Lauriola et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Paass \u0026amp; Giesselbach, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). These technologies have continuously iterated and upgraded at an astonishing pace, not only leading the forefront of AI research but also triggering transformative waves across multiple industries, including medical field (Ali et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), healthcare (Burns et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Xu \u0026amp; Wang, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), finance (Puchakayala, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), media (Br\u0026uuml;ns \u0026amp; Mei\u0026szlig;ner, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Feng et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), and manufacturing (Ghobakhloo et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), profoundly changing the way people work and live.\u003c/p\u003e\u003cp\u003eThese transformations across various sectors have not only driven industry efficiency and innovation but also prompted the education field to actively embrace generative artificial intelligence technology (Arinushkina, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Bozkurt \u0026amp; Sharma, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Sa\u0026uacute;de et al., \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). As a critical link in talent cultivation and social development, education is facing unprecedented opportunities and challenges brought by artificial intelligence (Bobula, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Wu et al., \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Particularly in the realm of academic writing, generative AI technology, with its groundbreaking advancements, is reshaping the academic landscape (Lorenc-Kukula, \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Mahdi \u0026amp; Alkhateeb, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Mondal, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Since the Transformer architecture revolutionized natural language processing, large-scale pre-trained models such as the GPT series and LLaMA have undergone continuous iterations, giving rise to a diverse range of writing assistance tools.\u003c/p\u003e\u003cp\u003eFor instance, GenAI-assisted writing tools assist students in conducting literature reviews by identifying pertinent scholarly articles (Behrooz et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; van Mossel et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2025\u003c/span\u003e), supplying contextual information related to writing topics (Chichekian \u0026amp; Benteux, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), summarizing key content (Behrooz et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), and generating personalized suggestions based on individual preferences and search patterns (Chichekian \u0026amp; Benteux, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Additionally, an GenAI-driven chatbot can support students in creating outlines for writing assignments (Kim et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). By improving the structure of their writing and creating a more interactive learning environment, the AI fosters effective writing habits. Through targeted questions and prompts, the chatbot encourages critical thinking, offering opportunities for reflection and enhancing self-regulation in the writing process (Lin \u0026amp; Chang, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Furthermore, AI tools, such as ChatGPT, support users in brainstorming ideas and receiving feedback on their writing, thereby enhancing the overall writing process (Atlas, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In addition, GenAI tools are increasingly recognized as valuable research assistants, capable of generating ideas, synthesizing complex information, and summarizing large volumes of text. These functions help researchers analyze data more efficiently and improve the quality of their academic writing (Berg, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe widespread application of generative AI technology in academic writing has not only transformed the way academic writing is conducted but also sparked varied attitudes and perspectives toward this emerging technology. Scholars and researchers exhibit a broad range of attitudes toward the integration of generative artificial intelligence in academic writing (Cohen \u0026amp; Moher, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Jin et al., \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2025\u003c/span\u003e; Pu et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Stojanov, \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2025\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eAn attitude is a person's comprehensive assessment of an object or topic, encompassing both emotional and evaluative aspects (Ajzen, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2001\u003c/span\u003e; Ajzen \u0026amp; Fishbein, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2000\u003c/span\u003e).When analyzing attitudes, the ABC (Affect\u0026ndash;Behavior\u0026ndash;Cognition) model is one of the most commonly used frameworks. It enables researchers to explore attitudes from three dimensions: affective (e.g., emotions such as trust or curiosity), behavioral (e.g., actual usage or intention to use), and cognitive (e.g., perceived usefulness or ease of use) (Breckler, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e1984\u003c/span\u003e; Jackson et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e1996\u003c/span\u003e; Millar \u0026amp; Tesser, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e1989\u003c/span\u003e; Wolff et al., \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). While this model provides a useful overarching framework, it tends to remain at a broad level of analysis and lacks the specificity needed to capture the nuanced impacts of generative AI in academic contexts.\u003c/p\u003e\u003cp\u003eMoreover, most existing studies focus heavily on the positive attitude (e.g., Atlas, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Behrooz et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; van Mossel et al., \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2025\u003c/span\u003e)\u0026mdash;while paying limited attention to negative perceptions. This imbalance may lead to an overly optimistic portrayal of generative AI adoption, overlooking critical barriers such as ethical concerns, mistrust, or anxiety, which could significantly influence actual user behavior and long-term integration in academic settings.\u003c/p\u003e\u003cp\u003eFurthermore, few studies have employed quantitative methods to rigorously verify or measure these negative attitudes. As a result, current understandings remain largely theoretical. The lack of empirical validation not only hinders the development of nuanced theoretical models but also weakens the foundation for evidence-based policymaking and practical interventions, potentially resulting in misaligned strategies that fail to address users\u0026rsquo; real concerns or resistance.\u003c/p\u003e\u003cp\u003eTo this end, this study aims to develop and validate a negative attitude scale toward GenAI-assisted academic writing among university students. Unlike general attitude measures, this scale is specifically designed for the academic writing context, ensuring high relevance and contextual sensitivity. Moreover, it shifts the research focus toward the often-overlooked negative perceptions, such as distrust, fear of over-reliance, and ethical concerns. By adopting a quantitative approach, the scale enables systematic measurement and comparison, providing robust empirical evidence for understanding the complexity of student responses to GenAI. This contributes not only to theory-building but also to the development of more balanced, responsive, and ethical educational policies and practices.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003eParticipants\u003c/h2\u003e\u003cp\u003eThe procedures used in this study adhere to the tenets of the Declaration of Helsinki. Prior to data collection, the ethical approval has been granted on 17/12/2024, by Macau Millennium College, with approval number MMCIRB-2024-002. The written consent form was embedded in the online questionnaire, and participants could proceed only after selecting the \"Willing to participate\" option. To protect participants' privacy, no identifiable information, such as names, contact details, or student ID, was collected. A total of 2,140 individuals accessed the survey, among whom 2,019 consented to participate and completed the questionnaire from March 26 to March 29, 2025. Respondents were from various regions, including China, such as Henan, Hebei, Beijing, Shanghai, Gansu, Guangxi, Heilongjiang, Jilin, Shaanxi, Chongqing, and Australia. The survey was distributed via the Wenjuanxing platform. The estimated completion time was approximately 3 minutes. A multi-stage design was adopted for instrument development and validation. The dataset was randomly divided into three groups. The first group (n\u0026thinsp;=\u0026thinsp;673) was used for item analysis and exploratory factor analysis. The second group (n\u0026thinsp;=\u0026thinsp;673) was used to conduct confirmatory factor analysis. The third group was (n\u0026thinsp;=\u0026thinsp;673) used for multi-group confirmatory factor analysis to examine measurement invariance across different subgroups. A detailed summary of the participants\u0026rsquo; demographic characteristics is provided in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003edemographic information\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"8\"\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\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\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eDemographic \u003c/p\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eGroup\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e\u003cp\u003eEFA Dataset\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e\u003cp\u003eCFA Dataset\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e\u003cp\u003eMG- CFA Dataset\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003eN\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eGender\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e108\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e16.00%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e141\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e21.00%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e160\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e23.80%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eFemale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e565\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e84.00%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e532\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e79.00%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e513\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e76.20%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eEducation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eUndergraduate\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e639\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e94.90%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e641\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e95.20%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e641\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e95.20%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eGraduate\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e34\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e5.10%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e32\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e4.80%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e32\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e4.80%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eBirthplace\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eUrban\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e201\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e29.90%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e196\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e29.10%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e189\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e28.10%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRural\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e472\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e70.10%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e477\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e70.90%\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e484\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e71.90%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eAnalytical procedure\u003c/h3\u003e\n\u003cp\u003eA multi-stage design was employed for instrument development and validation. The dataset was randomly partitioned into three subsets to support sequential psychometric analyses and ensure the robustness and generalizability of the findings. In the first stage, item analysis and exploratory factor analysis (EFA) were conducted using the first dataset. Item analysis included high-low group comparison, item-total correlation, and Cronbach\u0026rsquo;s alpha if item deleted, in order to identify items with low discrimination or internal inconsistency. Items that failed to show adequate discrimination between high and low scoring groups (i.e., with non-significant or weak differences) were considered to have insufficient discriminative power and were removed. Additionally, items with corrected item-total correlations below .30 or whose deletion improved the overall internal consistency were flagged for exclusion. Prior to EFA, the Kaiser\u0026ndash;Meyer\u0026ndash;Olkin (KMO) measure and Bartlett\u0026rsquo;s test of sphericity were used to confirm the data\u0026rsquo;s suitability for factor analysis (Kaiser, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e1974\u003c/span\u003e). EFA was then conducted and the number of factors was determined by parallel analysis with oblique rotation (promax). Parallel analysis was chosen because it compares the observed eigenvalues with those generated from random datasets to reduce the risk of factor over-extraction and improve the accuracy of factor retention decisions. Items with factor loadings below .40, cross-loadings above .30, or weak conceptual relevance were excluded to optimize the factor structure.\u003c/p\u003e\u003cp\u003eIn the second stage, confirmatory factor analysis (CFA) was conducted using the second dataset to validate the factor structure identified in EFA. Model fit was assessed using a combination of indices (Hu \u0026amp; Bentler, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e1999\u003c/span\u003e; McDonald \u0026amp; Ho, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2002\u003c/span\u003e): Comparative Fit Index (CFI) and Tucker\u0026ndash;Lewis Index (TLI) values\u0026thinsp;\u0026ge;\u0026thinsp;0.90 indicated acceptable fit, and \u0026ge;\u0026thinsp;0.95 indicated good fit. Root Mean Square Error of Approximation (RMSEA) values\u0026thinsp;\u0026le;\u0026thinsp;0.08 were considered acceptable (\u0026le;\u0026thinsp;0.06 for good fit), and Standardized Root Mean Square Residual (SRMR) values\u0026thinsp;\u0026le;\u0026thinsp;0.05 were also indicative of good model fit. Convergent validity was assessed by examining Average Variance Extracted (AVE) and Composite Reliability (CR); AVE\u0026thinsp;\u0026ge;\u0026thinsp;0.50 and CR\u0026thinsp;\u0026ge;\u0026thinsp;0.70 were considered acceptable (Hair et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2010\u003c/span\u003e). Discriminant validity was examined using the Fornell-Larcker criterion, ensuring that the square root of AVE for each construct exceeded the inter-construct correlations (Fornell \u0026amp; Larcker, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e1981\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eIn the third stage, multi-group confirmatory factor analysis (MG-CFA) was conducted using the third dataset to test measurement invariance across subgroups (e.g., gender, birthplace and educational background). A sequential procedure was employed to test the measurement invariance, including configural, metric, scalar, and strict invariance. Each step imposed increasing constraints on the model parameters. Changes in model fit between nested models were evaluated using ΔCFI, ΔTLI and ΔRMSEA, with ΔCFI and ΔTLI\u0026thinsp;\u0026le;\u0026thinsp;0.01 and ΔRMSEA\u0026thinsp;\u0026le;\u0026thinsp;0.015 indicating that invariance held at each level (Byrne et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e1989\u003c/span\u003e; Chen, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). This process ensured that the instrument measured the constructs equivalently across different groups.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003eStudy 1\u003c/h2\u003e\u003cdiv id=\"Sec7\" class=\"Section3\"\u003e\u003ch2\u003eItem development\u003c/h2\u003e\u003cp\u003eTo develop the initial pool of scale items, we conducted semi-structured interviews with five graduate students who had hands-on experience using generative AI tools to support academic writing. To ensure ethical standards and protect participant privacy, all interviewees were informed that the data collected would be used exclusively for academic purposes, with no personally identifiable information disclosed. The interview protocol focused on participants\u0026rsquo; attitudes toward AI-assisted writing, including their motivations for use, perceived benefits and challenges, shifts in writing practices, and views on authorship and identity. Each interview lasted approximately 40 minutes and was audio-recorded and transcribed. The transcripts were analyzed through a multi-stage coding process to ensure analytical rigor. During open coding, the researchers conducted line-by-line reading to extract initial concepts and assign descriptive labels to meaningful text segments. In axial coding, similar codes were grouped and refined to identify underlying patterns, while irrelevant or redundant content was eliminated. Selective coding was then used to synthesize these categories into three overarching themes\u0026mdash;language homogenization, thought outsourcing, and identity ambiguity\u0026mdash;that reflected participants\u0026rsquo; nuanced reflections on the implications of GenAI in academic writing and informed the item development. To enhance interpretive validity, the themes and key interpretations were sent back to the interviewees for member checking, allowing them to confirm accuracy or raise objections where necessary.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eContent validity\u003c/h2\u003e\u003cp\u003eSubsequently, to assess content validity, eight experts in education and information technology were invited to evaluate the relevance, clarity, and representativeness of each item on a 4-point Likert scale. The item-level content validity index (I-CVI) was calculated as the proportion of experts rating the item as either 3 or 4. Additionally, the scale-level content validity index (S-CVI) was computed as the average of all I-CVIs. As all items achieved an I-CVI of 0.80 or above\u0026mdash;commonly accepted as the minimum threshold for adequate content validity\u0026mdash;and the S-CVI exceeded 0.90, no revisions were deemed necessary at this stage.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eStudy 2\u003c/h3\u003e\n\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003eItem analysis\u003c/h2\u003e\u003cp\u003eIndependent samples \u003cem\u003et\u003c/em\u003e-tests were performed between the upper 27% group (total scores\u0026thinsp;\u0026ge;\u0026thinsp;64) and the lower 27% group (total scores\u0026thinsp;\u0026le;\u0026thinsp;48), revealing that all items differed significantly at the 95% confidence interval, thereby demonstrating strong discriminative validity. Furthermore, item-total correlations, measured by Pearson\u0026rsquo;s correlation coefficients, ranged from 0.742 to 0.901 (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.01), indicating robust associations between individual items and the overall scale score. The internal consistency reliability, assessed by Cronbach\u0026rsquo;s alpha, was 0.975. Deletion of any single item did not result in a significant increase in the alpha coefficient, confirming the high reliability and homogeneity of the instrument.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eExploratory factor analysis\u003c/h2\u003e\u003cp\u003eThe suitability of the data for exploratory factor analysis (EFA) was first evaluated. The Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy was exceptionally high at 0.965, indicating excellent factorability of the correlation matrix. Additionally, Bartlett\u0026rsquo;s test of sphericity was significant, χ\u0026sup2;(120)\u0026thinsp;=\u0026thinsp;15,019.689, p\u0026thinsp;\u0026lt;\u0026thinsp;0.001, rejecting the null hypothesis that the correlation matrix is an identity matrix and confirming the appropriateness of conducting factor analysis.\u003c/p\u003e\u003cp\u003eParallel analysis was then performed to determine the optimal number of factors to retain. Subsequently, an exploratory factor analysis was conducted employing an oblique Promax rotation, which allows for correlation among factors. Parallel analysis is preferred because it provides a more accurate and objective criterion by comparing the observed eigenvalues with those obtained from randomly generated data, thereby reducing the risk of over-extraction.\u003c/p\u003e\u003cp\u003eDuring the EFA process, items were considered for deletion based on the following criteria: low factor loadings (below 0.40), significant cross-loadings (loadings\u0026thinsp;\u0026ge;\u0026thinsp;0.30 on multiple factors), and lack of theoretical consistency with the intended construct. However, none of the items met these criteria for removal, and therefore, all original items were retained in the final factor structure. The analysis yielded a three-factor solution, which aligned well with the theoretically hypothesized factor structure.\u003c/p\u003e\u003cp\u003eThe extracted factors collectively explained a substantial proportion of the total variance as shown in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, and factor loadings demonstrated clear and interpretable item-factor associations consistent with the predefined dimensions.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eExploratory factor analysis results\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\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\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eItem\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLH\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTO\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eIA\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eCommunality\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eCronbach\u0026rsquo;s\u003c/p\u003e\u003cp\u003ealpha coefficient\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eVariance explained\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLH1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.912\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.818\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003e0.937\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003e34%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLH2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.881\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.786\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLH3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.864\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.832\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLH4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.768\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.731\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTO1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.881\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.826\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003e0.969\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003e29.90%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTO2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.884\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.873\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTO3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.755\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.786\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTO4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.796\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.847\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTO5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.807\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.86\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTO6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.756\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.856\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIA1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.931\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.878\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003e0.977\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003e20.30%\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIA2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.901\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.884\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIA3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.872\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.883\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIA4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.912\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.89\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIA5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.839\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.863\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eIA6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.856\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.874\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"7\"\u003e\u003cem\u003eNote: IH\u0026thinsp;=\u0026thinsp;language homogenization, TO\u0026thinsp;=\u0026thinsp;thought outsourcing and IA\u0026thinsp;=\u0026thinsp;identity ambiguity; Factor loadings below 0.40 are not displayed.\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eStudy 3: Confirmatory factor analysis\u003c/h2\u003e\u003cp\u003eTo test whether the observed variables adequately reflected the hypothesized latent constructs, confirmatory factor analysis (CFA) was conducted on the second dataset using the CB-SEM module in SmartPLS 4.0. The results in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e indicated a good model fit.\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\u003eModel fit statistics\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=\"left\" 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\u003cp\u003eχ2\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003edf\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eχ2/df\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eRMSEA 90% CI\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eGFI\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eSRMR\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003eTLI\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c9\"\u003e\u003cp\u003eCFI\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e382.217\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e101\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.01\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3.784\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e0.064[0.058\u0026ndash;0.071]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e0.932\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.016\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.978\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e0.981\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: RMSEA\u0026thinsp;=\u0026thinsp;Root Mean Square Error of Approximation 90% CI\u0026thinsp;=\u0026thinsp;90% Confidence Interval (for RMSEA) GFI\u0026thinsp;=\u0026thinsp;Goodness-of-Fit Index SRMR\u0026thinsp;=\u0026thinsp;Standardized Root Mean Square Residual TLI\u0026thinsp;=\u0026thinsp;Tucker-Lewis Index (also known as Non-Normed Fit Index, NNFI) CFI\u0026thinsp;=\u0026thinsp;Comparative Fit Index\u003c/em\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003eConvergent validity\u003c/h2\u003e\u003cp\u003eTo assess convergent validity, composite reliability (CR), average variance extracted (AVE), and standardized factor loadings were examined. As shown in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, all constructs demonstrated strong internal consistency, with CR values ranging from 0.938 to 0.975, exceeding the recommended threshold of 0.70. The AVE values ranged from 0.791 to 0.867, all above the minimum acceptable value of 0.50, indicating that each construct explains a substantial proportion of variance in its observed indicators. In addition, all standardized factor loadings were statistically significant and ranged from 0.875 to 0.945 (also see Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e), further supporting the convergent validity of the scale.\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\u003eCR and AVE results\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=\"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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFactor\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eComposite reliability\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAverage variance extracted\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eidentity ambiguity\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.975\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.867\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003elanguage homogenization\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.938\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.791\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ethought outsourcing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.973\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.855\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003eDiscriminant validity\u003c/h2\u003e\u003cp\u003eDiscriminant validity was assessed using the Fornell-Larcker criterion. As shown in Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, the square root of the AVE for each construct (shown on the diagonal) exceeded its correlations with the other constructs, indicating adequate discriminant validity.\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\u003eFornell-Larcker results\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=\"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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFactor\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1. identity ambiguity\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e\u003cb\u003e0.931\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2. language homogenization\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.689\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u003cb\u003e0.890\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3. thought outsourcing\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.887\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.699\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u003cb\u003e0.925\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003eStudy 4\u003c/h2\u003e\u003cdiv id=\"Sec16\" class=\"Section3\"\u003e\u003ch2\u003eMulti-group Confirmatory Factor Analysis\u003c/h2\u003e\u003cp\u003eThe multi-group confirmatory factor analysis (MG-CFA) was conducted using the third dataset to assess measurement invariance across birthplace, sex, and education level. Sequential tests of configural, metric, scalar, and residual invariance all demonstrated good model fit, with no significant declines when equality constraints were imposed. Fit indices as shown in Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e remained within acceptable thresholds (e.g., ΔCFI\u0026thinsp;\u0026lt;\u0026thinsp;0.01, ΔTFI\u0026thinsp;\u0026lt;\u0026thinsp;0.01, ΔRMSEA\u0026thinsp;\u0026lt;\u0026thinsp;0.015), supporting full measurement invariance across these demographic groups. These findings indicate that the measurement model operates equivalently across subgroups defined by birthplace, sex, and education level, allowing for meaningful group comparisons.\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\u003eMG-CFA results\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"14\"\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=\"left\" 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\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=\"char\" char=\".\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c14\" colnum=\"14\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDemographic Variable\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\u003eχ2\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003edf\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eχ2/df\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eΔχ2\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eΔ\u003cem\u003edf\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003ep\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c9\"\u003e\u003cp\u003eCFI\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c10\"\u003e\u003cp\u003eΔCFI\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c11\"\u003e\u003cp\u003eTLI\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c12\"\u003e\u003cp\u003eΔTLI\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c13\"\u003e\u003cp\u003eRMSEA\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c14\"\u003e\u003cp\u003eΔRMSEA\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"4\" rowspan=\"5\"\u003e\u003cp\u003eBirthplace\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eUnconstrained\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e658.548\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e202.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e3.260\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.970\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.964\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.058\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eConfigural invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e671.728\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e215.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e3.124\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e13.180\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e13.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.434\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.970\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e0.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.966\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e0.002\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.056\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e-0.002\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMetric invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e688.453\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e231.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.980\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e16.725\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e16.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.404\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.970\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e0.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.969\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e0.003\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.054\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e-0.002\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eScalar invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e694.789\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e237.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.932\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e6.336\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e6.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.387\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.970\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e0.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.969\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e0.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.054\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e0.000\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eResidual invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e755.253\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e253.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.985\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e60.464\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e16.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.967\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e-0.003\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.969\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e0.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.054\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e0.000\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"4\" rowspan=\"5\"\u003e\u003cp\u003eSex\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eUnconstrained\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e600.549\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e202\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.973\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.973\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.968\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.054\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eConfigural invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e608.032\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e215\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.828\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e7.483\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.876\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.974\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.971\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e0.003\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.052\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e-0.002\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMetric invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e661.399\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e231\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.863\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e53.367\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e16\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.971\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e-0.003\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.97\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e-0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.053\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eScalar invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e695.124\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e237\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.933\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e33.725\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.969\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e-0.002\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.969\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e-0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.054\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eResidual invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e787.328\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e253\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e3.112\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e92.204\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e16\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.964\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e-0.005\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.966\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e-0.003\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.056\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e0.002\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"4\" rowspan=\"5\"\u003e\u003cp\u003eEducation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eUnconstrained\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e580.597\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e202.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.874\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.975\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.970\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.053\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eConfigural invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e591.988\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e215.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.753\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e11.391\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e13.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.578\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.975\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e0.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.972\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e0.002\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.051\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e-0.002\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMetric invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e603.929\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e231.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.614\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e11.941\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e16.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.748\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.975\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e0.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.974\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e0.002\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.049\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e-0.002\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eScalar invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e606.879\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e237.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.561\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e2.950\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e6.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0.815\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.976\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.975\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.048\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e-0.001\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eResidual invariance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e737.815\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e253.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e2.916\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e130.936\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e16.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c9\"\u003e\u003cp\u003e0.968\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c10\"\u003e\u003cp\u003e-0.008\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c11\"\u003e\u003cp\u003e0.970\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c12\"\u003e\u003cp\u003e-0.005\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c13\"\u003e\u003cp\u003e0.053\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c14\"\u003e\u003cp\u003e0.005\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we adopted a multi-study design to develop and validate a scale measuring university students\u0026rsquo; attitudes toward the application of generative artificial intelligence (Generative AI) in academic writing. A total of 2,019 participants were recruited, providing a large and representative sample. Based on preliminary interviews and qualitative analysis, three core dimensions were identified through coding and thematic extraction: language homogenization (4 items), thought outsourcing (6 items), and identity ambiguity (6 items). These dimensions comprehensively capture students\u0026rsquo; attitudes toward the potential impacts of generative AI on academic writing, from the perspectives of language style, cognitive processes, and self-identity.\u003c/p\u003e\u003cp\u003eSubsequently, item analysis and exploratory factor analysis (EFA) were conducted. Parallel analysis confirmed the appropriateness of a three-factor structure, which aligned closely with the theoretically hypothesized dimensions. All items showed strong loadings and clear structural distinctions within their respective factors, with no items needing to be removed. Further confirmatory factor analysis (CFA) provided robust support for the structural model, demonstrating good model fit as well as significant convergent and discriminant validity, thereby confirming the construct validity of the scale. In addition, multi-group confirmatory factor analysis (MG-CFA) indicated that the scale exhibited structural invariance across subgroups defined by birthplace, gender, and educational level, confirming its suitability for meaningful cross-group comparisons and its broad applicability across diverse demographic profiles.\u003c/p\u003e\u003cp\u003eThe language homogenization factor captures students\u0026rsquo; perceptions of how generative AI contributes to the stylistic standardization and emotional flattening of academic writing. The items under this dimension reflect a clear pattern of linguistic convergence, including the frequent adoption of AI-recommended passive constructions, complex syntactic patterns, and formalized paragraph structures. Additionally, participants expressed a tendency to use high-frequency, generic vocabulary suggested by AI, which diminishes their ability to adapt language to diverse topics or express personal voice. Importantly, the avoidance of subjective tone and emotional expression\u0026mdash;as noted in one item\u0026mdash;underscores a shift toward neutral, impersonal academic discourse, making writing appear emotionally flat or detached. Together, these items consistently reflect a core concern: that excessive reliance on generative AI may gradually erode the individuality, creativity, and affective richness of academic writing, leading to stylistic homogenization across authors and contexts.\u003c/p\u003e\u003cp\u003eThe thought outsourcing dimension is similarly well represented, with items reflecting a progressive delegation of core cognitive processes\u0026mdash;such as ideation, critical reasoning, and coherence checking\u0026mdash;to generative AI. This factor captures not only behavioral reliance (e.g., default use of AI-generated arguments or structure) but also a deeper psychological disengagement from academic authorship, including diminished agency, reduced creativity, and a sense of detachment from one\u0026rsquo;s own thought process.\u003c/p\u003e\u003cp\u003eLastly, items under the identity ambiguity factor reflect students\u0026rsquo; growing uncertainty and internal conflict regarding their academic roles and authorship legitimacy. These items articulate the blurred boundaries between human and AI contributions in academic writing, including concerns about intellectual ownership, ethical detachment, and the performative nature of authorship in an AI-mediated academic environment. The metaphorical expressions used by participants\u0026mdash;such as \u0026ldquo;ghostwriter,\u0026rdquo; \u0026ldquo;academic tailor,\u0026rdquo; or \u0026ldquo;Turing Test\u0026rdquo; participant\u0026mdash;not only highlight the depth of this identity confusion but also resonate strongly with contemporary discourse on authorship and AI ethics.\u003c/p\u003e\u003cdiv id=\"Sec18\" class=\"Section2\"\u003e\u003ch2\u003eImplications\u003c/h2\u003e\u003cp\u003eThis study advances the theoretical understanding of attitudes toward technology by moving beyond the generalist Affect\u0026ndash;Behavior\u0026ndash;Cognition (ABC) model. Rather than focusing solely on affective reactions or behavioral intentions, we propose a domain-specific framework consisting of language homogenization, thought outsourcing, and identity ambiguity\u0026mdash;three dimensions that capture the cognitive, linguistic, and identity-related dilemmas students face when engaging with generative AI in academic writing. This situated approach provides a more nuanced and explanatory account of user attitudes, highlighting how AI reshapes students\u0026rsquo; expressive style, cognitive autonomy, and sense of authorship. Compared to general models that assess broad sentiments like willingness to use or perceived usefulness, our framework offers deeper insight into the psychological and behavioral transformations emerging in AI-integrated academic contexts.\u003c/p\u003e\u003cp\u003eFrom an educational perspective, the scale provides actionable insights for educators and institutions seeking to responsibly integrate generative AI into academic settings. By identifying specific challenges\u0026mdash;such as linguistic standardization, diminished critical engagement, and blurred authorship\u0026mdash;educators can better anticipate the unintended consequences of AI-assisted writing. These findings underscore the need for pedagogical strategies that go beyond teaching students how to use AI tools, and instead focus on when, why, and to what extent they should be used. Curriculum designers might incorporate critical AI literacy, reflective writing tasks, and authorship ethics into writing instruction to help students retain cognitive ownership and identity coherence. The scale also serves as a diagnostic tool for monitoring students\u0026rsquo; evolving attitudes and guiding adaptive teaching interventions in rapidly changing technological landscapes.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003eLimitations and suggestions\u003c/h2\u003e\u003cp\u003eAlthough this study employed rigorous psychometric procedures\u0026mdash;including exploratory and confirmatory factor analyses\u0026mdash;to establish the structural validity of the scale, future research could benefit from methodological triangulation. For instance, applying item response theory (IRT) could provide more detailed insights into individual item functioning across different subgroups, while network analysis might reveal latent inter-item relationships and centrality patterns. These complementary approaches would offer a more nuanced understanding of the scale\u0026rsquo;s dimensional structure and enhance the robustness of its construct validity.\u003c/p\u003e\u003cp\u003eThe current study is cross-sectional in nature, which constrains our ability to examine the developmental trajectory of students\u0026rsquo; attitudes toward generative AI in academic writing. Given that attitudes may evolve as students gain more exposure, experience, or institutional guidance regarding AI technologies, future research should adopt longitudinal designs. Tracking attitude changes over time would help determine the stability or transformation of the identified dimensions (e.g., whether thought outsourcing intensifies or diminishes with AI familiarity), and provide a dynamic view of student adaptation in AI-mediated learning environments.\u003c/p\u003e\u003cp\u003eWhile the scale effectively captures attitudinal dimensions, its predictive validity remains to be established. Specifically, it is unclear how these attitudes\u0026mdash;such as concerns over identity ambiguity or tendencies toward thought outsourcing\u0026mdash;translate into actual academic writing behaviors or outputs. Future studies should examine the extent to which these factors predict writing performance indicators such as originality, coherence, ethical conduct, or engagement with feedback. Such investigations would not only validate the scale\u0026rsquo;s practical utility but also offer educators diagnostic insights for intervention and support.\u003c/p\u003e\u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThis study developed and validated a scale measuring university students\u0026rsquo; attitudes toward the use of generative AI in academic writing, encompassing three key dimensions: language homogenization, thought outsourcing, and identity ambiguity. The scale demonstrated strong reliability and validity across diverse groups. These findings contribute to a deeper theoretical understanding and offer practical guidance for educators in designing responsible and reflective AI-integrated writing instruction.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003ch2\u003eEthical approval\u003c/h2\u003e\u003cp\u003eThe procedures used in this study adhere to the tenets of the Declaration of Helsinki. Prior to data collection, the ethical approval has been granted by Macau Millennium College, with approval number MMCIRB-2024-002.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eInformed consent\u003c/strong\u003e\u003cp\u003eThe written consent form was embedded in the online questionnaire, and participants could proceed only after selecting the \"Willing to participate\" option. To protect participants' privacy, no identifiable information, such as names, contact details, or student ID, was collected. Participants were assured of anonymity and confidentiality, with the understanding that their responses would be used solely in aggregated form for academic publications and presentations. They were also informed that they could withdraw from the study at any time, and in such cases, their data would be securely destroyed upon request.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003eDeclaration\u003c/p\u003e\u003cp\u003eThis study received no funding.\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eAuthor ContributionsConceptualization: YYC and SZMethodology: YYC and CTXSoftware: YYC and CTXValidation: YYC and TZHFormal analysis: YYC and SZData curation: YYC and SZWriting \u0026ndash; original draft preparation: YYCWriting \u0026ndash; review and editing: YXX and TZHVisualization: YXXSupervision: YYCProject Administration: YYC\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eData is provided within the manuscript or supplementary information files\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAjzen I (2001) Nature and Operation of Attitudes. Ann Rev Psychol 52(1):27\u0026ndash;58. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1146/annurev.psych.52.1.27\u003c/span\u003e\u003cspan address=\"10.1146/annurev.psych.52.1.27\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAjzen I, Fishbein M (2000) Attitudes and the Attitude-Behavior Relation: Reasoned and Automatic Processes. Eur Rev Social Psychol 11(1):1\u0026ndash;33. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/14792779943000116\u003c/span\u003e\u003cspan address=\"10.1080/14792779943000116\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAli H, Rehmani MH, Shāh SZ (2023) \u0026amp; service), S. (Online. \u003cem\u003eAdvances in Deep Generative Models for Medical Artificial Intelligence\u003c/em\u003e (1st ed. 20). Springer Nature Switzerland. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-031-46341-9\u003c/span\u003e\u003cspan address=\"10.1007/978-3-031-46341-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eArinushkina AA (2025) Integration strategies of generative AI in higher education. Integration strategies of generative artificial intelligence in higher education. IGI Global\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAtlas S (2023) \u003cem\u003eChatGPT for higher education and professional development: A guide to conversational AI\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://digitalcommons.uri.edu/cba_facpubs/548\u003c/span\u003e\u003cspan address=\"https://digitalcommons.uri.edu/cba_facpubs/548\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBehrooz H, Lipizzi C, Korfiatis G, Ilbeigi M, Powell M, Nouri M (2023) Towards Automating the Identification of Sustainable Projects Seeking Financial Support: An AI-Powered Approach. Sustainability 15(12):9701. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/su15129701\u003c/span\u003e\u003cspan address=\"10.3390/su15129701\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBerg C (2023) \u003cem\u003eThe case for generative AI in scholarly practice\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2139/ssrn.4407587\u003c/span\u003e\u003cspan address=\"10.2139/ssrn.4407587\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBobula M (2024) Generative artificial intelligence (AI) in higher education: a comprehensive review of challenges, opportunities, and implications. \u003cem\u003eJournal of Learning Development in Higher Education\u003c/em\u003e, \u003cem\u003e30\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.47408/jldhe.vi30.1137\u003c/span\u003e\u003cspan address=\"10.47408/jldhe.vi30.1137\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBozkurt A, Sharma RC (2024) Transforming education with generative AI. In \u003cem\u003eTransforming education with generative artificial intelligence\u003c/em\u003e. IGI Global, Engineering Science Reference. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4018/979-8-3693-1351-0\u003c/span\u003e\u003cspan address=\"10.4018/979-8-3693-1351-0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBreckler SJ (1984) Empirical validation of affect, behavior, and cognition as distinct components of attitude. J Personal Soc Psychol 47(6):1191\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBr\u0026uuml;ns JD, Mei\u0026szlig;ner M (2024) Do you create your content yourself? Using generative artificial intelligence for social media content creation diminishes perceived brand authenticity. J Retailing Consumer Serv 79:103790. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jretconser.2024.103790\u003c/span\u003e\u003cspan address=\"10.1016/j.jretconser.2024.103790\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBurns B, Nemelka B, Arora A (2024) Practical implementation of generative artificial intelligence systems in healthcare: A United States perspective Practical implementation of generative artificial intelligence systems in healthcare: A United States perspective. Future Healthc J 11(3):100166\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eByrne BM, Shavelson RJ, Muth\u0026eacute;n B (1989) Testing for the equivalence of factor covariance and mean structures: The issue of partial measurement invariance. Psychol Bull 105(3):456\u0026ndash;466. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/0033-2909.105.3.456\u003c/span\u003e\u003cspan address=\"10.1037/0033-2909.105.3.456\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen FF (2007) Sensitivity of goodness of fit indexes to lack of measurement invariance. Struct Equation Modeling: Multidisciplinary J 14(3):464\u0026ndash;504\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChichekian T, Benteux B (2022) The potential of learning with (and not from) artificial intelligence in education. Front Artif Intell 5:903051. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3389/frai.2022.903051\u003c/span\u003e\u003cspan address=\"10.3389/frai.2022.903051\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCohen JF, Moher D (2025) Generative artificial intelligence and academic writing: friend or foe? J Clin Epidemiol 179:111646. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jclinepi.2024.111646\u003c/span\u003e\u003cspan address=\"10.1016/j.jclinepi.2024.111646\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFeng W, Li Y, Ma C, Yu L (2025) From ChatGPT to Sora: Analyzing Public Opinions and Attitudes on Generative Artificial Intelligence in Social Media. IEEE Access 13:14485\u0026ndash;14498. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/ACCESS.2025.3530683\u003c/span\u003e\u003cspan address=\"10.1109/ACCESS.2025.3530683\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFornell C, Larcker DF (1981) Evaluating structural equation models with unobservable variables and measurement error. J Mark Res 18(1):39\u0026ndash;50. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2307/3151312\u003c/span\u003e\u003cspan address=\"10.2307/3151312\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGhobakhloo M, Fathi M, Iranmanesh M, Vilkas M, Grybauskas A, Amran A (2024) Generative artificial intelligence in manufacturing: opportunities for actualizing Industry 5.0 sustainability goals. J Manuf Technol Manage 35(9):94\u0026ndash;121. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/JMTM-12-2023-0530\u003c/span\u003e\u003cspan address=\"10.1108/JMTM-12-2023-0530\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHair J, Black W, Babin B, Anderson R (2010) \u003cem\u003eMultivariate data analysis\u003c/em\u003e (7th Editio). Prentice-Hall, Inc\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHu LT, Bentler PM (1999) Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria Versus New Alternatives. Struct Equation Modeling: Multidisciplinary J 6(1):1\u0026ndash;55. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/10705519909540118\u003c/span\u003e\u003cspan address=\"10.1080/10705519909540118\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJackson LA, Hodge CN, Gerard DA, Ingram JM, Ervin KS, Sheppard LA (1996) Cognition, Affect, and Behavior in the Prediction of Group Attitudes. Personal Soc Psychol Bull 22(3):306\u0026ndash;316. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1177/0146167296223009\u003c/span\u003e\u003cspan address=\"10.1177/0146167296223009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJin F, Sun L, Pan Y, Lin C-H (2025) High heels, compass, spider-man, or drug? Metaphor analysis of generative artificial intelligence in academic writing. Computers Educ 228:105248. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.compedu.2025.105248\u003c/span\u003e\u003cspan address=\"10.1016/j.compedu.2025.105248\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKaiser HF (1974) An index of factorial simplicity. Psychometrika 39(1):31\u0026ndash;36. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/BF02291575\u003c/span\u003e\u003cspan address=\"10.1007/BF02291575\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKim J, Yu S, Detrick R, Li N (2025) Exploring students\u0026rsquo; perspectives on Generative AI-assisted academic writing. Educ Inform Technol 30(1):1265\u0026ndash;1300. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10639-024-12878-7\u003c/span\u003e\u003cspan address=\"10.1007/s10639-024-12878-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLauriola I, Lavelli A, Aiolli F (2022) An introduction to Deep Learning in Natural Language Processing: Models, techniques, and tools. Neurocomputing (Amsterdam) 470:443\u0026ndash;456. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.neucom.2021.05.103\u003c/span\u003e\u003cspan address=\"10.1016/j.neucom.2021.05.103\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLin MP-C, Chang D (2020) Enhancing Post secondary Writers\u0026rsquo; Writing Skills with a Chatbot: A Mixed-Method Classroom Study. Educational Technol ༆ Soc 23(1):78\u0026ndash;92. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.30191/ETS.202001_23(1).0006\u003c/span\u003e\u003cspan address=\"10.30191/ETS.202001_23(1).0006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLorenc-Kukula K (2025) Cutting-edge AI tools revolutionizing scientific research in life sciences. Biotechnologia (Poznan) 106(1):77\u0026ndash;102. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5114/bta/200803\u003c/span\u003e\u003cspan address=\"10.5114/bta/200803\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMahdi HS, Alkhateeb A (2025) Revolutionising Essay Evaluation: A Cutting-Edge Rubric for AI-Assisted Writing. Int J Computer-Assisted Lang Learn Teach 15(1):1\u0026ndash;19. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4018/IJCALLT.368226\u003c/span\u003e\u003cspan address=\"10.4018/IJCALLT.368226\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcDonald RP, Ho M-HR (2002) Principles and practice in reporting structural equation analyses. Psychol Methods 7(1):64\u0026ndash;82. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/1082-989X.7.1.64\u003c/span\u003e\u003cspan address=\"10.1037/1082-989X.7.1.64\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMillar MG, Tesser A (1989) The effects of affective-cognitive consistency and thought on the attitude-behavior relation. J Exp Soc Psychol 25(2):189\u0026ndash;202. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/0022-1031(89)90012-7\u003c/span\u003e\u003cspan address=\"10.1016/0022-1031(89)90012-7\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMondal H (2025) The Future of Writing: How Artificial Intelligence is Shaping the Way We Write. Indian J Vascular Endovascular Surg 12(1):74\u0026ndash;75. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4103/ijves.ijves_160_24\u003c/span\u003e\u003cspan address=\"10.4103/ijves.ijves_160_24\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePaass G, Giesselbach S (2023) \u003cem\u003eFoundation Models for Natural Language Processing: Pre-trained Language Models Integrating Media\u003c/em\u003e (1st ed. 20). Springer International Publishing. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-031-23190-2\u003c/span\u003e\u003cspan address=\"10.1007/978-3-031-23190-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePu Z, Shi C, Jeon CO, Fu J, Liu S, Lan C, Yao Y, Liu Y, Jia B (2024) ChatGPT and generative AI are revolutionizing the scientific community: A Janus-faced conundrum. IMeta 3(2):e178\u0026ndash;n. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e/a\u003c/span\u003e\u003cspan address=\"http:///a\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/imt2.178\u003c/span\u003e\u003cspan address=\"10.1002/imt2.178\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePuchakayala PRA (2024) Generative Artificial intelligence Applications in Banking and Finance sector. World J Adv Res Reviews 23(1):3105\u0026ndash;3120. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.30574/wjarr.2024.23.1.1999\u003c/span\u003e\u003cspan address=\"10.30574/wjarr.2024.23.1.1999\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eReddy P, Chand K, Sharma K, Sharma B, Sharma S (2025) Evolution of Generative Artificial Intelligence: A Review of the Developed and Developing. Eng Sci. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.30919/es1529\u003c/span\u003e\u003cspan address=\"10.30919/es1529\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSa\u0026uacute;de S, Barros JP, Almeida I (2024) Impacts of Generative Artificial Intelligence in Higher Education: Research Trends and Students\u0026rsquo; Perceptions. Social Sci (Basel) 13(8):410. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/socsci13080410\u003c/span\u003e\u003cspan address=\"10.3390/socsci13080410\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eStojanov A (ed) (2025) Editors\u0026rsquo; Newsroom: Polishing our words or replacing our voices? Generative artificial intelligence in academic writing. \u003cem\u003eSocial Behavior and Personality\u003c/em\u003e, \u003cem\u003e53\u003c/em\u003e(3), 1\u0026ndash;2. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2224/sbp14842\u003c/span\u003e\u003cspan address=\"10.2224/sbp14842\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003evan Mossel S, Oude-Wolcherink MJ, de Feria Cardet RE, de Geus-Oei L-F, Vriens D, Koffijberg H, Saing S (2025) Artificial Intelligence as a New Research Ally? Performing AI-Assisted Systematic Literature Reviews in Health Economics: AI as an Ally? Performing AI-Assisted Systematic Literature Reviews in Health Economics. PharmacoEconomics 43(6):647\u0026ndash;650. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s40273-025-01481-4\u003c/span\u003e\u003cspan address=\"10.1007/s40273-025-01481-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWolff K, Nordin K, Brun W, Berglund G, Kvale G (2011) Affective and cognitive attitudes, uncertainty avoidance and intention to obtain genetic testing: An extension of the Theory of Planned Behaviour. Psychol Health 26(9):1143\u0026ndash;1155. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/08870441003763253\u003c/span\u003e\u003cspan address=\"10.1080/08870441003763253\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWu D, Chen L, Han Y (2023) Challenges that generative artificial intelligence poses to higher education and management and the countermeasures\u0026mdash; A visualized analysis of literature from 2019 to 2023 using CiteSpace. 733\u0026ndash;738. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1145/3660043.3660174\u003c/span\u003e\u003cspan address=\"10.1145/3660043.3660174\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXu R, Wang Z (2024) Generative artificial intelligence in healthcare from the perspective of digital media: Applications, opportunities and challenges. Heliyon, 10(12), e32364\u003c/span\u003e\u003c/li\u003e\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":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"attitude, Generative AI-assisted academic writing, scale development, scale validation","lastPublishedDoi":"10.21203/rs.3.rs-6974623/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6974623/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study adopted a multi-stage research design to develop and validate a scale measuring university students\u0026rsquo; attitudes toward the application of generative artificial intelligence (Generative AI) in academic writing. Based on preliminary interviews and qualitative analysis, three core dimensions were identified: language homogenization, thought outsourcing, and identity ambiguity, reflecting the potential impacts of AI on language style, cognitive processes, and authorship identity, respectively. A total of 2,019 participants were recruited. During scale development, item analysis and exploratory factor analysis (EFA) were first conducted. Parallel analysis confirmed the appropriateness of a three-factor structure. Subsequent confirmatory factor analysis (CFA) demonstrated good model fit, as well as strong convergent and discriminant validity. Finally, multi-group confirmatory factor analysis (MG-CFA) further supported the structural invariance of the scale across subgroups defined by gender, birthplace, and educational level, indicating its broad applicability. Despite its solid reliability, validity and broad applicability, the scale still has limitations, such as the need for longitudinal studies, predictive validity testing, and validation using multiple methods. Overall, the scale enriches the understanding of user attitudes toward technology beyond the traditional Affective\u0026ndash;Behavioral\u0026ndash;Cognitive model, and offers concrete educational insights for the responsible integration of generative AI in academic writing.\u003c/p\u003e","manuscriptTitle":"Development and Validation of a Scale Assessing College Students’ Negative Attitudes Toward Generative AI-Assisted Academic Writing","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-30 07:17:35","doi":"10.21203/rs.3.rs-6974623/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-01-06T17:41:02+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-26T11:44:51+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-03T13:07:07+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"283357228753288363216195051454074577849","date":"2025-11-02T12:44:48+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"273248214694746924187548943659274291987","date":"2025-11-02T12:00:29+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-09-18T08:49:43+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-18T08:48:22+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-09-18T06:54:17+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-04T08:56:57+00:00","index":"","fulltext":""},{"type":"submitted","content":"Humanities and Social Sciences Communications","date":"2025-09-04T08:51:04+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"ac04a70f-44a8-44fb-ba65-8119ccdd1403","owner":[],"postedDate":"September 30th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"in-revision","subjectAreas":[{"id":55451361,"name":"Social science/Education"},{"id":55451362,"name":"Social science/Science technology and society"}],"tags":[],"updatedAt":"2026-05-13T22:38:43+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-30 07:17:35","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6974623","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6974623","identity":"rs-6974623","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}