Control Discrepancy and Anxiety Symptomatology: A Distinct Metacognitive Vulnerability Beyond Traditional Control Constructs

Control Discrepancy and Anxiety Symptomatology: A Distinct Metacognitive Vulnerability Beyond Traditional Control Constructs | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 23 June 2025 V1 Latest version Share on Control Discrepancy and Anxiety Symptomatology: A Distinct Metacognitive Vulnerability Beyond Traditional Control Constructs Authors : Christopher J. Davis 0009-0005-2859-5354 [email protected] and Sydnie R. Spearman Authors Info & Affiliations https://doi.org/10.22541/au.175067477.74668572/v1 475 views 166 downloads Contents Abstract Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Objectives: Control beliefs are well-established predictors of anxiety symptomatology, yet existing models emphasize the protective effects of high perceived control without considering the influence of individual differences in metacognitive appraisals of control—namely, the evaluation of the control one believes to have and the control they desire to have. In the present study, we propose that control discrepancy (i.e., the subjective appraisal of the misalignment between perceived and desired control) is a distinct metacognitive vulnerability factor for generalized anxiety disorder (GAD). Methods: In a large non-clinical sample (N = 1,325; M age = 35.9, SD = 16.0), we examined whether control discrepancy predicted elevated GAD risk and greater GAD symptom severity beyond traditional control constructs. Participants completed validated measures assessing mental health symptoms and control beliefs. Results: Results from LASSO-regularized logistic and generalized ordered logistic regressions revealed that control discrepancy was a significant predictor of both GAD risk and symptom severity. Notably, perceiving more control than desired presented a particularly elevated risk, even at high levels of perceived control. These associations remained robust after adjusting for depressive symptoms, perceived stress, perceived control, desired control, and their interaction. Conclusion: Findings provide initial evidence of control discrepancy as a distinct control belief that is uniquely associated with GAD symptomatology beyond that of perceived control, desired control, and their interaction. These results call for a re-evaluation of interventions that primarily target increasing perceived control and underscore the clinical potential of tailoring treatment to reduce control discrepancy, particularly when perceived control greatly exceeds desired control. math_shortcuts Abstract Objectives: Control beliefs are well-established predictors of anxiety symptomatology, yet existing models emphasize the protective effects of high perceived control without considering the influence of individual differences in metacognitive appraisals of control—namely, the evaluation of the control one believes to have and the control they desire to have. In the present study, we propose that control discrepancy (i.e., the subjective appraisal of the misalignment between perceived and desired control) is a distinct metacognitive vulnerability factor for generalized anxiety disorder (GAD). Methods: In a large non-clinical sample (N = 1,325; M age = 35.9, SD = 16.0), we examined whether control discrepancy predicted elevated GAD risk and greater GAD symptom severity beyond traditional control constructs. Participants completed validated measures assessing mental health symptoms and control beliefs. Results: Results from LASSO-regularized logistic and generalized ordered logistic regressions revealed that control discrepancy was a significant predictor of both GAD risk and symptom severity. Notably, perceiving more control than desired presented a particularly elevated risk, even at high levels of perceived control. These associations remained robust after adjusting for depressive symptoms, perceived stress, perceived control, desired control, and their interaction. Conclusion: Findings provide initial evidence of control discrepancy as a distinct control belief that is uniquely associated with GAD symptomatology beyond that of perceived control, desired control, and their interaction. These results call for a re-evaluation of interventions that primarily target increasing perceived control and underscore the clinical potential of tailoring treatment to reduce control discrepancy, particularly when perceived control greatly exceeds desired control. Keywords : Control beliefs; Perceived Control; Control Discrepancy; Anxiety; GAD-7 Control Discrepancy and Anxiety Symptomatology: A Distinct Metacognitive Vulnerability Beyond Traditional Control Constructs Control beliefs are central to models of psychological functioning. In the context of generalized anxiety disorder (GAD), individuals often have trouble controlling worry over life outcomes, which has been consistently linked to both symptom presence and severity. This disruption directly relates to perceived control—the belief in one’s ability to influence outcomes—where low control is a known risk factor for anxiety (Amoura et al., 2014). Indeed, contemporary models (e.g., Barlow’s triple vulnerability model) posit that diminished perceived control over internal emotions and external stressors creates a generalized risk for developing anxiety disorders (Barlow, 2004). Conversely, high perceived control is typically associated with better coping and less anxiety, and dominant treatment modalities often aim to increase one’s sense of control, which has been linked to symptom improvement (Gallagher et al., 2014b). However, control beliefs involve not only how much control one has (or feels they have) but also how much control one desires (i.e., desired control). This is particularly noteworthy, as previous work has demonstrated that desired control varies across individuals and contexts, with some exhibiting a strong desire to direct most outcomes, while others may be more comfortable relinquishing control in specific situations. Thus, the interplay between perceived and desired control may be crucial for understanding anxiety, yet it has been relatively overlooked. Perhaps more critically, an area that has yet to be explored is one’s subjective appraisal of the degree of misalignment between their perceived and desired control, which we refer to in the present paper as control discrepancy. This construct asks: is the control you believe that you have over your life outcomes less than, equal to, or more than the control you ideally want? Intuitively, when perceived control falls short of desired control (e.g., feeling unable to control important outcomes despite wishing for a high degree of control) anxiety may increase. This idea aligns with longstanding theories that lacking a sense of control over stressors tend to increase anxiety and feelings of helplessness (Seligman, 1975). Yet an equally compelling question is whether the reverse could also be problematic: could perceiving to have more control than one desires contribute to anxiety or distress? Most research on anxiety and control has focused on control deficits, implicitly discrediting the potentially negative correlates of perceiving excess control relative to the amount of control one desires. In the present paper, we examine control discrepancy as a potentially important, yet understudied, dimension of control beliefs that may elucidate how insufficient or excess perceived control may be associated with increased GAD risk and symptom severity. Control Beliefs in Diagnostic Criteria and Assessment of GAD The relevance of control to the experience of anxiety is not only theoretical but also evident in the most utilized diagnostic frameworks and assessment tools used to evaluate GAD. Indeed, diagnostic criteria of GAD according to The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5; APA, 2013) include: “excessive anxiety and worry (apprehensive expectation)” and “the individual finds it difficult to control the worry.” Additional diagnostic criteria for GAD according to the DSM-5 encompasses a range of somatic and cognitive symptoms, including “restlessness”, “irritability”, and “difficulty concentrating”, and “muscle tension” which are symptoms that often reflect underlying efforts to manage or gain a sense of control over perceived threats (Urben et al., 2024). In both clinical practice and research, GAD is frequently assessed using the Generalized Anxiety Disorder-7 (GAD-7; Spitzer et al., 2006) which is a brief, psychometrically robust screening measure highlighted for its reliability and sensitivity to treatment effects (Löwe et al., 2008). The GAD-7 includes items that implicitly implicate control, such as “not being able to stop or control worrying,” “feeling afraid, as if something awful might happen,” and “worrying too much about different things.” These items suggest disruptions in perceived control underlie GAD symptoms, aligning with research on perceived control as a transdiagnostic anxiety factor (Gallagher et al., 2014b). Specifically, meta-analytic findings reveal a strong negative association between perceived control and anxiety, with the most pronounced effect observed in participants with generalized anxiety disorder. Moreover, Gallagher and colleagues, (2014a) provided longitudinal evidence suggesting that increases in perceived control during cognitive-behavioral therapy (CBT) are associated with corresponding decreases in anxiety symptoms, and that perceived control mediates treatment outcomes regardless of specific diagnostic category. Together, these insights suggest that perceived control is not only relevant in the diagnostic framework and assessment of GAD, but also an important mechanism driving psychological outcomes. Low Perceived Control as a Dominant Focus and the Neglected Risk of High Perceived Control Longstanding theoretical models and empirical work have also nominated low perceived control as a fundamental cognitive feature of psychological distress (Lazarus & Folkman, 1984) and vulnerability to anxiety disorders (Gallagher et al., 2014b). This perspective is rooted in theories such as Seligman’s learned helplessness theory (1975), which proposed that low perceived control over aversive outcomes may increase one’s vulnerability to anxiety and depression. In this vein, Chorpita and Barlow (1998) proposed that individuals high in trait anxiety tend to interpret ambiguous situations as threatening due to an underlying sense of uncontrollability (i.e., low perceived control). Empirical studies provide evidence supporting these theoretical underpinnings, suggesting that individuals who report low perceived control over life events are more likely to experience elevated anxiety symptoms, particularly under conditions of uncertainty or stress (Thompson et al., 2001). Moreover, experimental studies have repeatedly demonstrated that perceived uncontrollability over aversive stimuli predicts increased physiological arousal and psychological distress (Bhanji et al., 2014; Müller, 2012), including anxiety symptoms (Bollini et al., 2004; Davis et al., 2025). Thus, low perceived control has been largely viewed as a robust cognitive factor in the development and persistence of anxiety symptomatology. While extant literature predominately emphasizes low perceived control as a key risk factor for anxiety, less attention has been given to the potential drawbacks of excessive control and its contribution to anxiety symptomatology. Though often protective, excessive perceived control may be maladaptive in some contexts. For instance, hypervigilance, a common feature of GAD, may reflect an excessive effort to monitor and enhance feelings of control within a situation or environment in order to avoid perceived threats (Bar-Haim et al., 2007). While these behaviors may initially serve a protective function (e.g., intentionally engaging in behaviors aimed at increasing perceived control to decrease downstream anxiety symptoms), they may ultimately contribute to the maintenance of anxiety (Richards et al., 2014), even when a high degree of control over one’s environment is endorsed. Moreover, excessive control has been implicated in maladaptive perfectionism (Soenens et al., 2005), which is associated with hallmark GAD symptoms (e.g., worry and rumination) as well as a heightened vulnerability to anxiety disorders (Egan et al., 2011). These instances challenge the dominant assumption that high perceived control necessarily begets better psychological outcomes. To be clear, we do not argue that the benefits of heightened perceived control are invalid, nor are the risks associated with low perceived control overstated. Instead, we argue that more nuance is needed within the literature—specifically pertaining to when the magnitude of one’s perceived control may pose an increased risk for GAD symptomatology and factors that may inform adaptive versus maladaptive expressions of control beliefs. Extending this argument, we maintain that high perceived control is beneficial so long as one does not subjectively appraise their perceived control to outweigh their desire for it. Person-Environment Fit Theory and Control Discrepancy A well-established framework for understanding the complexities of control beliefs on mental health and psychological well-being is the person-environment (P-E) fit theory (French Jr. et al., 1982). This theory posits that well-being is optimized when there is congruence between personal characteristics (e.g., needs, desires) and environmental affordances (e.g., opportunities, demands). In the domain of control, misalignment between desired and perceived control—which we refer to in the present paper as control discrepancy —may be particularly relevant for understanding emotional responses and mental health outcomes. In perhaps the most direct test of this idea, Conway and colleagues (1992) found that control discrepancies—whether having more or less control than desired—were associated with greater psychological strain, including higher negative affect and lower quality of life, supporting a U-shaped relationship between control discrepancy and emotional outcomes. Later research by Myles and colleagues (2020) as well as Moulding and colleagues (2007, 2008) also examined the relationship between perceived and desired control on mental health symptoms and found evidence that low perceived control coupled with high desired control (i.e., insufficient control) was associated with greater distress. It is worth noting that neither Myles and colleagues (2020) nor Moulding and colleagues (2008) modeled the U-shaped relationship hypothesized and found by Conway and colleagues (1992), limiting their ability to directly examine the presence of psychological distress simultaneously for insufficient and excess control discrepancies. The Limitations of Derived Control Discrepancy Scores and the Case for Subjective Appraisal To date, most studies examining the relationship between control discrepancy and psychological outcomes have utilized mathematically derived proxies, rather than the individuals’ own appraisal of their control discrepancy (i.e., utilizing a difference score or modeling the interaction between desired and perceived control). This approach implicitly assumes that individuals’ measured discrepancy scores are equal to their subjective assessment of the presence or magnitude of their control discrepancy. Though preferable to ignoring the construct entirely, this approach is insufficient in capturing the subjective experience of control discrepancy, thus limiting the inferences drawn from these studies. For instance, an individual who scores high on both perceived and desired control may show no numerical control discrepancy, but still subjectively appraise one to exist. In this case, control discrepancy as a subjective appraisal could serve as a more robust predictor of one’s psychological distress than their derived score. This notion is in line with findings in related domains that demonstrate both (1) the lack of agreement between subjective appraisals of a construct and more objective measures (e.g., self-rated health status vs. objective health status; socioeconomic status vs. income or education), as well as (2) their divergent outcomes. A well-documented example of this phenomenon is provided by Jylhä (2009), who found that self-rated health often diverges significantly from objective clinical assessments—yet these subjective ratings possess unique predictive value. For instance, individuals’ perception of their health status has been shown to independently predict future disability, declines in functional ability, and mental health outcomes, beyond what was explained by clinical indicators. Similarly, Idler and Benyamini (1997) reported that subjective health ratings predict mortality and healthcare utilization even after accounting for objective health indicators. These findings underscore the distinct importance of subjective appraisals in psychological outcomes and give credence to the possibility that individuals’ perception of their control discrepancy may be a more meaningful predictor of mental health than mathematically derived scores alone. This issue may be especially relevant in the context of GAD, as it is characterized by pervasive cognitive biases and distorted information processing (Beck & Clark, 1997). Thus, subjective control discrepancy could represent a more ecologically valid index of the cognitive-affective processes underlying GAD symptoms—one that better aligns with the distorted perceptions often observed in anxious individuals. In line with this, most contemporary cognitive models of anxiety maintain that individuals with anxiety oftentimes demonstrate heightened sensitivity to perceived threats and exaggerated appraisals of potential harm, which may play a critical role in distorting perceptions of control (see Beck & Clark, 1997). Specific to control beliefs, an individual with GAD may perceive a significant control discrepancy where none demonstrably exists or may drastically overestimate their level of control discrepancy relative to their mathematically derived discrepancy score. This highlights a critical measurement and theoretical challenge, such that derived scores may fail to capture the subjective experience of control discrepancies (i.e., how discrepant one feels ), especially in the case that the appraisals are distorted by anxiety-related cognitive biases. To this extent, cognitive models of anxiety also emphasize that biased cognitions typical of anxious individuals include intolerance of uncertainty (Dugas et al., 2004) and inflated responsibility beliefs (Salkovskis, 1996), both of which are central to the maintenance of symptoms and may exacerbate distress independently of objective control conditions. Thus, relying solely on perceived control or mathematically derived control discrepancy scores risk underestimating the clinical significance that assessing one’s subjective appraisal of their control discrepancy may offer. Present Study Building on the theoretical significance of control beliefs in anxiety symptomatology, the present study investigates the association between control discrepancy and GAD risk and symptom severity as assessed by the GAD-7. The study is guided by two primary hypotheses. First, (H1) we hypothesize that increased control discrepancy is associated with increased odds of meeting the GAD-7 cutoff. We expect to find a nonlinear association, such that increased control discrepancies (i.e., perceiving much more control than desired or perceiving much less control than desired) are linked to elevated anxiety risk. Additionally, (H2) we hypothesize that increased control discrepancy is associated with an increased likelihood of heightened GAD symptom severity. In examining these hypotheses, we aim to clarify whether subjective control discrepancies—both in the form of insufficient control and excess control—are meaningfully associated with GAD risk and symptom severity beyond traditional constructs such as perceived control, desired control, and their interaction. math_shortcuts Method Participants. A non-clinical adult sample of N = 1,416 ( M age = 35.9, SD = 16.0) were recruited through Prolific Academic (www.prolific.com), an online platform designed to connect researchers with participants who receive compensation for participating in research studies. Inclusion criteria included participants being at least 18 years of age. Exclusion criteria included failure to obtain a ReCaptcha score of at least .7 on Qualtrics (Griffen et al., 2021) and failure to fully complete measures for the variables of interest. After enforcing exclusion criteria, a final analytic sample of N = 1,325 remained. 428 participants (32.30%) reported anxiety symptoms that met the recommended GAD-7 cutoff line (i.e., x GAD-7 score ≥ 10; Spitzer et al., 2006). According to Spitzer et al. (2006), GAD severity scores were distributed as follows: 534 participants (40.30%) reported minimal symptoms, 363 participants (27.40%) reported mild symptoms, 229 participants (17.28%) reported moderate symptoms, and 199 participants (15.02%) reported severe symptoms. The final analytic sample comprised 54.79% identified their gender as woman, 44.68% as man, and 0.53% as non-binary. Measure Anxiety Symptoms. Anxiety symptoms were assessed through the GAD-7 (Spitzer et al., 2006). The GAD-7 is a 7-item self-report measure in which respondents rate the frequency of anxiety symptoms experienced over the previous two weeks. The GAD-7 is scored on a 4-point Likert scale ranging from 0 (“not at all”) to 3 (“nearly every day”) and scores range from 0-21. The GAD-7 demonstrated adequate internal consistency in the present study (α = .95). Control Discrepancy. Given no existing multidimensional measure, we developed a face-valid single-item measure of control discrepancy. This item was based off widely used single-item measures of subjective psychological constructs including the Single-Item Self-Esteem scale (SISE; Robins, Hendin, & Trzesniewski, 2001), Self-Rated Health scale (SRH; Idler & Benyamini, 1997), and the Single-Item Life Satisfaction scale (S-SWLS; Cheung & Lucas, 2014). The item used to measure control discrepancy reads: “In general, how much control do you believe to have over your life outcomes” with answer choices ranging from 1 (“extremely less than I desire”) to 9 (“extremely more than I desire”), both corresponding to a high magnitude of control discrepancy. The middlemost answer choice 5 (“as much as I desire”) corresponds to no control discrepancy. Depressive Symptoms. Depressive symptoms were assessed through the Beck Depression Inventory II (BDI-II; Beck et al., 1996). The BDI-II is a 21-item self-report measure in which respondents rate the frequency of experienced depressive symptoms over the previous two weeks. It is scored on a 4-point Likert scale ranging from 0 to 3 and scores ranging from 0-63. In the present study, the BDI-II displayed adequate internal consistency (α = .94). Perceived Control. Perceived control was assessed using the Sense of Control Scale (SOCS; Lachman & Weaver, 1998). The SOCS is a twelve item, self-report measure which includes two subscales to measure distinct domains of perceived control: personal mastery and perceived constraints. The SOCS is scored on a 7-point Likert scale ranging from 1 (“strongly disagree”) to 7 (“strongly agree”). In the present study, the SOCS displayed sufficient internal consistency (α = .88). Desired Control. Desired control was assessed using the Desirability of Control Scale (DCS; Burger & Cooper, 1979). The DCS is a twenty item self-report measure which includes 3 subscales: leadership, self-control, and others-control. The DCS is scored on a 7-point Likert scale ranging from 1 (“does not apply to me at all”) to 7 (“always applies to me”). In the present study the DCS demonstrated sufficient internal consistency in each sample (α = .80). Perceived Stress. Perceived stress was assessed using the Perceived Stress Scale-4 (PSS-4; Cohen & Williamson, 1988). The PSS-4 is a 4-item self-report measure where respondents report their thoughts and feelings about stressors in their lives over the past month. The PSS-4 is scored on a 5-point Likert scale ranging from 0 (“never”) to 4 (“very often”). The total score is calculated by summing all items. Scores range from 0-20 with higher scores indicating higher levels of PS. In the present study, the PSS-4 demonstrated acceptable internal consistency (α = .74). Demographic Characteristics. Participants were asked to indicate the following demographic information: age, race/ethnic background, gender identity, and socioeconomic status. Procedure Participants first read and signed a consent form approved by the [REDACTED FOR REVIEW] Institutional Review Board (IRB). The consent form provided information about study procedures. Participants then completed a battery of surveys. Lastly, participants were compensated for their time through a nominal cash payment. Preregistration & Data Analytic Strategy The present study draws upon data from a larger series of studies conducted by our research group. All procedures received institutional approval from the [REDACTED FOR REVIEW] IRB. Additionally, the present study was preregistered prior to data analysis on OSF (osf.io/k7z83). Supplementary material, analytic code, and raw data files are present in the study repository. Data analysis was conducted in Stata/BE 19.5 (StataCorp, 2023). Significance levels were set to α = .05 for all hypothesis testing. Standardized parameter estimates and coefficients were reported to mitigate multicollinearity concerns and to enhance interpretability across predictors. To ensure robust parameter estimates, bootstrapped standard errors (5,000 replications) were estimated for each model. LASSO Regularization. To determine a parsimonious and theoretically coherent set of predictors for inclusion in the logistic and ordinal logistic regression models, we employed a Least Absolute Shrinkage and Selection Operator (LASSO) regularized logistic regression using the Extended Bayesian Information Criterion (EBIC). Ten-fold cross-validation was used to select the optimal regularization parameter (λ) in order to minimize prediction error while reducing overfitting. This approach allowed for data-driven covariate selection while addressing issues of multicollinearity and overfitting. The final set of predictors retained by the LASSO procedure was then used in all subsequent hypothesis-testing models. Although LASSO is often used for prediction, the primary aim of this study is explanatory rather than classificatory. As such, model performance was examined across thresholds primarily to assess the stability and interpretability of findings, rather than to optimize classification metrics. Predicted probabilities were dichotomized at a threshold of 0.35, such that values ≥ 0.35 were considered indicative of probable GAD cases. While slightly improved specificity was observed at a threshold of 0.45, this difference was minimal (see Supplementary Figure S1). Thus, a 0.35 threshold was ultimately chosen to reflect the model’s explanatory goals. H1. A hierarchical binary logistic regression (including four blocks) with odds ratio was conducted to assess H1 and examine the added utility of control discrepancy beyond other model covariates. Block 1 included perceived control and desired control; Block 2 included Block 1 covariates and the interaction between perceived control and desired control; Block 3 included Block 2 covariates and depressive symptoms, perceived stress, age, gender identification (reference = woman); Block 4 included Block 3 covariates and control discrepancy as a linear, quadratic, cubic, and quartic term. A series of Likelihood-Ratio tests were conducted to determine the best fitting model. In addition to statistical significance, model performance was assessed using classification diagnostics including sensitivity, specificity, accuracy, and area under the receiver operating characteristic (ROC) curve to gauge the model’s discriminative capacity and predictive accuracy (Hosmer et al., 2013). Predicted GAD risk probabilities were estimated to illustrate GAD risk across observed levels of control discrepancy at different levels of perceived and desired control. H2. An ordered logistic regression with odds ratio was estimated to examine the relationship between control discrepancy and the likelihood of reporting GAD-7 scores in four symptom severity categories (i.e., minimal, mild, moderate, and severe). The final model included control discrepancy, depressive symptoms, age, gender identification (reference = women), and desired control, perceived control, and the interaction between desired x perceived control as linear terms and perceived stress as linear and quadratic terms. When assessing model assumptions, the proportional odds assumption was found to be violated. To address this, a generalized ordered logistic regression model with autofitting via the autofit option in Stata’s (2025) gologit2 command was estimated. This approach relaxes the proportional odds assumption only for variables that violate it, while maintaining it for others. Predicted probabilities for severe GAD symptom categorization were estimated to illustrate probabilities of severe GAD across observed levels of control discrepancy at different levels of perceived control and desired control. Results Table 1 reports demographic characteristics (i.e., age, gender identity, and race/ethnicity) and descriptive statistics for each continuous variable of interest. Figure 1 illustrates the raw distribution of control discrepancy scores for the final analytic sample by anxiety cutoff classification. Model Selection Using LASSO-Regularized Logistic Regression LASSO Regularization. The LASSO path evaluated 50 values of the regularization parameter, with the optimal model selected at λ = 4.79, corresponding to the minimum EBIC value of EBIC = 971.06 and a pseudo- R ² of .39. Based on the LASSO model, predictors retained for inclusion in the final model to examine H1 were as follows: depressive symptoms, perceived stress, perceived control, desired control, perceived control x desired control, age, gender (at 3 levels: woman (reference), man, neither man nor woman), and control discrepancy (cubic). H1: Control discrepancy is associated with increased odds of meeting the GAD-7 cutoff. To assess the incremental utility of control discrepancy, we conducted a hierarchical logistic regression model comprising four blocks (see Supplementary Table S1 for model fit indices and parameter estimates for each block). Block 4—the only block that included control discrepancy— provided the best fit to the data (Wald χ²(11) = 353.30, p < .001; pseudo- R ² = .57), confirmed by an LR-test comparing Block 3 and Block 4 (χ²(3) = 34.58, p < .001). The Block 4 model indicated excellent discriminative capacity: AUC = 0.95; SE = 0.01, 95% CI [0.94, 0.96] (see Figure 2 for illustration of AUC). Additionally, the Block 4 model correctly classified 86.27% of cases (PPV = 73.89%; NPV = 92.50%), with a sensitivity of 83.33% and specificity of 87.48%, suggesting a strong balance between detecting true cases of GAD and minimizing false positives. In Block 4, we found control discrepancy (cubic) was significantly associated with increased GAD risk (OR = 1.18, SE = 0.07, z = 2.79, p = .005), such that every unit increase in control discrepancy corresponded with an 18.20% increased risk of meeting the GAD-7 cutoff. Notably, the interaction between perceived control x desired control remained significant in Block 4 (OR = 0.65, SE = 0.07, z = -3.99, p < .001). This provides evidence of distinctiveness between control discrepancy and the interaction between perceived control x desired control, as the inclusion of control discrepancy improved model fit and explained more variance in GAD risk. Additionally, control discrepancy and perceived control x desired control demonstrated opposite associations, with control discrepancy increasing one’s GAD risk and perceived control x desired control lowering one’s GAD risk. To further contextualize the unique contribution of control discrepancy beyond other control constructs, we examined the likelihood of meeting the GAD-7 cutoff across different levels of perceived control (i.e., low = M - 1σ; high = M + 1σ) and desired control (i.e., low = M - 1σ; high = M + 1σ). Average marginal effects revealed that increases in control discrepancy were significantly associated with higher predicted probabilities of meeting the GAD-7 cutoff across all combinations of perceived and desired control:\(\frac{\text{dy}}{\text{dx}}\ \) range = 4.66% (high perceived control, high desired control) – 5.45% (low perceived control, high desired control), all p ’s < .001. These effects remained robust regardless of whether individuals reported high or low levels of perceived or desired control, underscoring the unique and consistent relationship between control discrepancy and GAD risk. Next, we assessed the contribution of control discrepancy at specific values across combinations of perceived control and desired control (see Figure 3). While the combination of low perceived control and high desired control (compared to other combinations) was associated with higher GAD risk across levels of control discrepancy, individuals endorsing a control discrepancy of M control discrepancy + 2σ consistently showed elevated risk of meeting the GAD-7 cutoff compared to other levels of control discrepancy — even at high levels of perceived control. This pattern suggests that the magnitude of control discrepancy is strongly associated with increased GAD risk irrespective of perceived control, particularly when one appraises their perceived control to greatly exceed their desired control. To further clarify the differential GAD risk associated with heightened control discrepancy, we conducted an exploratory pairwise comparison analysis of the GAD-7 risk between participants who responded to the control discrepancy item with “extremely more [control] than I desire” compared to every other control discrepancy response (see Table 2). We found that those who perceived to have “extremely more [control] than I desire” presented a significantly higher risk of meeting the GAD-7 cutoff compared to any other control discrepancy response. This suggests that individuals who perceive themselves to have extremely more control than they desire may represent a distinct high-risk group for GAD, with significantly elevated symptom likelihood compared to all other control discrepancy profiles. H2: Control discrepancy is associated with increased GAD symptom severity. Next, we examined the association between control discrepancy and GAD symptom severity through a generalized ordered logistic regression. A Likelihood Ratio (LR) test suggested that the best fitting H2 model contained each variable included in the Block 4 model of H1, except control discrepancy as a quadratic and cubic term (LR χ²(2) = 0.96, p = .620). Instead, the LR test suggested that the H2 model was best specified when including control discrepancy only as a linear term. The final H2 model demonstrated good overall fit and explained a moderate portion of variance in GAD symptom severity (LR χ²(11) = 1049.42, pseudo- R ² = .32). Results indicated a consistent, positive association between control discrepancy and GAD symptom severity across each severity category (OR = 1.20, SE = 0.08, z = 2.74, p = .006). Specifically, for each one-unit increase in control discrepancy, the odds of reporting greater GAD symptom severity increased by approximately 20%. Notably, the proportional odds test for the control discrepancy variable was nonsignificant ( p = .474), indicating that the association between control discrepancy and GAD symptom severity did not significantly vary across outcome thresholds.11Although the proportional odds assumption was satisfied for control discrepancy, it was not satisfied for depressive symptoms, necessitating the use of generalized ordered logistic regression. The interaction between perceived control x desired control was also significantly associated with GAD symptom severity (OR = 0.86, SE = 0.05, z = -2.58, p = .010), such that the protective effect of perceived control weakened as desired control increased.22The proportional odds test for the interaction term of perceived control x desired control was also nonsignificant ( p = .562). Next, we estimated the marginal effect of control discrepancy on predicted probabilities of GAD symptom severity at different combinations of perceived and desired control. Results revealed that each unit increase in control discrepancy was associated with a 2.28% decrease in the probability of minimal GAD symptom severity ( SE = 0.84%, z = –2.72, p = .007), alongside increases of 0.76% and 1.36% in the probabilities of moderate and severe GAD symptom severity, respectively ( zs > 2.50, ps < .013). However, control discrepancy had no marginal effect on mild anxiety was not statistically significant ( z = 1.66, p = .098). These effects were consistent across combinations of high and low levels of perceived and desired control. To further probe the effects of control discrepancy on GAD symptom severity, we lastly estimated predicted probabilities of being classified with severe GAD symptoms across standardized values of control discrepancy, from −2σ to +2σ. As illustrated in Figure 4, increases in control discrepancy were associated with a consistent and substantial rise in the probability of severe GAD symptoms, underscoring its independent contribution to symptom severity. While this pattern was evident across all participants, predicted probabilities were also stratified by the interaction between perceived control and desired control to contextualize the overall risk profile. Notably, individuals characterized by low perceived control and high desired control demonstrated the highest overall risk of severe anxiety across all levels of control discrepancy. In contrast, those with high perceived control and low desired control exhibited a comparatively lower and more stable risk trajectory. Discussion The present study examined the association between control discrepancy—defined as the subjective appraisal of misalignment between one’s perceived and desired control—and GAD risk and symptom severity, as assessed by the GAD-7. Our findings highlight increased GAD risk linked to control discrepancy, especially when perceived control exceeds desired control. Below, we discuss key findings, their implications, and potential directions for future research are discussed below. math_shortcuts Summary of Key Findings In partial support of H1, we found that control discrepancy is associated with elevated GAD risk. That is, participants who reported more control than desired were more likely to reach the suggested GAD-7 clinical threshold. However, this association was asymmetric , with respondents who perceived to have substantially more control than desired demonstrating a steeper GAD risk. Furthermore, participants who reported having “extremely more control than I desire” exhibited the highest odds of meeting the GAD-7 cutoff compared to other control discrepancy responses. Notably, this finding was consistent across levels of perceived and desired control. For instance, the combination of low perceived control and high desired control consistently predicted higher GAD risk than any other pairing—which is consistent with past research. However, this was specific to one’s level of control discrepancy, as increased values of control discrepancy (i.e., going from less control than desired to more control than desired) were associated with greater GAD risk across combinations of perceived and desired control. This suggests that one’s level of control discrepancy: (1) may not necessarily reflect their independent levels of perceived and desired control, nor the statistically derived mismatch of both constructs; and (2) may represent a unique psychological vulnerability contributing to anxiety symptoms. Along these lines, the present findings emphasize the independent association of control discrepancy in GAD risk and symptom severity, beyond not only perceived and desired control, but other known risk factors such as depressive symptoms, perceived stress, age, and gender identification. Moreover, our findings demonstrate that even at high levels of perceived control, large control discrepancies (especially when perceiving excess control) continue to predict elevated GAD risk, suggesting that discrepancies between what individuals want and what they perceive they have in terms of control may be more central to understanding anxiety outcomes than previously recognized. Theoretical and Clinical Implications The findings in the present study are novel and were garnered through a cross-sectional design, which precludes causal inferences, and should be interpreted as such. This notwithstanding, our findings provide a promising direction for future research into how one’s subjective appraisal of the discrepancy between their control beliefs—not a statistically derived score of discrepancy—may contribute to anxiety vulnerability. First, our results extend the traditional deficit model of control in anxiety disorders, which posits that low perceived control is a primary driver of distress (Gallagher et al., 2014a). Instead, they lend preliminary support to a subjective discrepancy model of control, in which both excessive and insufficient perceived control—relative to one’s desired amount—can independently contribute to anxiety symptomatology. From a theoretical perspective, these findings are consistent with control mismatch frameworks, which propose that psychological distress arises from discrepancies between actual and desired control (Moulding et al., 2007). Though previous studies have demonstrated findings (i.e., the additive effect of perceived and desired control were positively associated with mental health symptoms), they failed to measure individuals’ subjective appraisal of their control discrepancy but instead utilized a statistically derived ‘mismatch’ (Moulding et al., 2006; Miles et al., 2020) or discrepancy score (Conway et al., 1992). Thus, our findings underscore the importance of considering not only discrepancies between independent control beliefs, but one’s own perception of control discrepancies. Clinically, these preliminary findings emphasize the value of reassessing interventions that focus primarily on enhancing perceived control. For individuals with elevated control discrepancy—particularly those who perceive to have more control than desired—therapeutic approaches might be more effective if they focus on helping clients regulate their desire for control, adjust expectations, or relinquish unnecessary control. Though aspects of Dialectical Behavioral Therapy (DBT) and CBT encourage individuals to challenge cognitive distortions and understand when to assert and relinquish control, further adaptations to these approaches may be deemed necessary if these findings hold within clinical populations. Moreover, these findings underscore the importance of assessing subjective appraisals of control discrepancy, rather than relying on evaluating perceived control, alone. If these findings hold, the inclusion of brief, direct assessments of control discrepancy (such as the single-item measure used in this study) may assist clinicians in broadly identifying individuals at elevated risk for anxiety—even among those who report high levels of perceived control. Nevertheless, future replication in diverse and longitudinal samples is an essential next step to determine the reliability and broader applicability of these findings before they can be fully translated into clinical practice. Limitations and Future Directions In addition to the limitations mentioned above, additional limitations warrant consideration and may provide a valuable impetus for future research. First, the use of a single-item measure for control discrepancy—though pragmatic and face-valid—limits the ability to assess multidimensional aspects of this construct. To date, there is no formal multidimensional measure of control discrepancy, nor a hypothesized multidimensional structure. Future research should prioritize the development of a clear theoretical framework outlining the conceptual and dimensional structure of control discrepancy and its relationship with anxiety symptomatology. Such theoretical frameworks should address: (1) the potential dimensions or domains in which control discrepancy manifests (e.g., emotional, behavioral, cognitive); (2) the mechanisms through which control discrepancy contributes to anxiety symptomatology; and (3) individual differences that may moderate the effects of control discrepancy (e.g., coping style, trait control beliefs, or intolerance of uncertainty). Addressing these initial conceptual questions is critical for grounding future empirical work in a robust and systematic theoretical framework. Additionally, the development and validation of a multi-item scale to capture nuances in control discrepancy, including domain-specific discrepancies (e.g., occupational, interpersonal, and intrapersonal), is necessary for advancing this line of research. Second, while the sample was large and demographically diverse, it was also non-clinical, which limits generalizability to clinical populations. Though a large proportion of the sample reported clinically distressing anxiety symptoms, it is likely that the findings, particularly regarding the association between control discrepancy and anxiety severity categorization, may differ in a clinical population. Future research should replicate these findings in diagnosed GAD samples or longitudinally assess how control discrepancy changes with treatment. In this vein, it is worth reiterating that the cross-sectional nature of the data utilized in the present study precludes any conclusions about causality. It is plausible that anxiety symptoms influence perceptions of control discrepancy (e.g., anxious individuals may be more susceptible to perceiving control discrepancies), or that both are influenced by a third variable (e.g., maladaptive perfectionism or intolerance of uncertainty). Future research should prioritize longitudinal and experimental approaches to clarify the temporal dynamics of this relationship. Finally, the asymmetric risk curve associated with perceiving more control than desired should be explored further, as it is unclear why these associations were stronger compared to those found for individuals perceiving less control than desired. Qualitative studies or ecological momentary assessment (EMA) approaches could clarify the real-time psychological mechanisms behind the increased GAD risk and symptom severity associated with excessive control. Despite these limitations, the analytic rigor employed in the present study serves as a methodological strength. We employed advanced statistical techniques, including LASSO-regularized logistic regression with EBIC and bootstrapped standard errors, to enhance model parsimony, parameter stability, and ensure that our observed relationships were not artifacts of overfitting or multicollinearity. In demonstrating consistent, nonlinear, and asymmetric effects across a large and demographically diverse sample, our results offer a stable empirical foundation from which future longitudinal and experimental studies can extend. Conclusion The present study provides compelling preliminary evidence that control discrepancy—not just low perceived control—is a robust predictor of both GAD risk and symptom severity. Our findings suggest that individuals reporting heightened control discrepancies, particularly those perceiving more control than desired, present an increased GAD risk and severity, which extends traditional models and suggests the need for a more nuanced understanding of control in anxiety and related disorders. Clinicians and researchers alike may benefit from integrating the concept of control discrepancy into theoretical models to better support individuals navigating the complexities of anxiety. math_shortcuts References Amoura, C., Berjot, S., Gillet, N., & Altintas, E. (2014). Desire for control, perception of control: Their impact on autonomous motivation and psychological adjustment. Motivation and Emotion, 38 (3), 323–335. DOI: 10.1007/s11031-013-9379-9. Averill, J. R. (1973). Personal control over aversive stimuli and its relationship to stress. Psychological bulletin , 80 (4), 286. Bandura, A. (2001). Social cognitive theory: An agentic perspective. Annual review of psychology , 52 (1), 1-26. Barlow, D. H. (2002). Anxiety and its disorders: The nature and treatment of anxiety and panic (2nd ed.). New York: Guilford Press. Barlow, D. H. (2004). Anxiety and its disorders: The nature and treatment of anxiety and panic . Guilford press. Beck, A. T., & Clark, D. A. (1997). An information processing model of anxiety: Automatic and strategic processes. Behaviour Research and Therapy, 35 (1), 49–58. https://doi.org/10.1016/S0005-7967(96)00069-1 Beck, A. T., Steer, R. A., & Brown, G. K. (1996). Manual for the Beck Depression Inventory-II . San Antonio, TX: Psychological Corporation. Bhanji, J. P., & Delgado, M. R. (2014). Perceived control influences neural responses to setbacks and promotes persistence. Neuron , 83 (6), 1369-1375. Bollini, A. M., Walker, E. F., Hamann, S., & Kestler, L. (2004). The influence of perceived control and locus of control on the cortisol and subjective responses to stress. Biological psychology, 67 (3), 245-260. Burger, J. M. (1989). Negative reactions to increases in perceived personal control . Journal of Personality and Social Psychology, 56 (2), 246–256. Burger, J. M., & Cooper, H. M. (1979). The desirability of control. Motivation and Emotion, 3 (4), 381–393. DOI: 10.1007/BF00994052 Cheung, F., & Lucas, R. E. (2014). Assessing the validity of single-item life satisfaction measures: Results from three large samples. Quality of Life research , 23 , 2809-2818. Cohen, S., & Williamson, G. M. (1988). Perceived stress in a probability sample of the United States. In S. Spacapan & S. Oskamp (Eds.), The Social Psychology of Health (pp. 31–67). Newbury Park, CA: Sage. Conway, T. L., Vickers Jr, R. R., & French Jr, J. R. (1992). An application of person—Environment fit theory: Perceived versus desired control. Journal of Social Issues, 48(2), 95-107. French, J. R. P., Caplan, R. D., & Harrison, R. V. (1982). The mechanisms of job stress and strain. Chichester, England: Wiley. Gallagher, M. W., Bentley, K. H., & Barlow, D. H. (2014). Perceived control and vulnerability to anxiety disorders: A meta-analytic review. Cognitive Therapy and Research, 38 (5), 571–584. https://doi.org/10.1007/s10608-014-9624-x Gallagher, M. W., Naragon-Gainey, K., & Brown, T. A. (2014). Perceived control is a transdiagnostic predictor of cognitive-behavior therapy outcome for anxiety disorders. Cognitive Therapy and Research, 38 (1), 10–22. DOI: 10.1007/s10608-013-9587-3. Halford, C., Jonsdottir, I. H., & Eek, F. (2012). Perceived stress, psychological resources and salivary cortisol. The role of saliva cortisol measurement in health and disease , 67-86. Idler, E. L., & Benyamini, Y. (1997). Self-rated health and mortality: a review of twenty-seven community studies. Journal of health and social behavior, 21-37. Jordet, G., Gemser, M. E., & Lemmink, K. P. (2006). Perceived control and anxiety. International. Journal of Sport Psychology , 37 , 281-298. Jylhä, M. (2009). What is self-rated health and why does it predict mortality? Towards a unified conceptual model. Social science & medicine, 69(3), 307-316. Lachman, M. E. (2006). Perceived control over aging-related declines: Adaptive beliefs and behaviors. Current Directions in Psychological Science, 15 (6), 282-286. Lachman, M. E., & Weaver, S. L. (1998). The sense of control as a moderator of social class differences in health and well-being. Journal of Personality and Social Psychology , 74 (3), 763–773. Lazarus, R. S., and Folkman, S. (1984). Stress, Appraisal, and Coping . New York: Springer. Löwe, B., Decker, O., Müller, S., Brähler, E., Schellberg, D., Herzog, W., & Herzberg, P. Y. (2008). Validation and standardization of the Generalized Anxiety Disorder Screener (GAD-7) in the general population. Medical care , 46 (3), 266-274. Mineka, S., & Hendersen, R. W. (1985). Controllability and predictability in acquired motivation. Annual review of psychology . Moulding, R., & Kyrios, M. (2007). Desire for control, sense of control and obsessive-compulsive symptoms. Cognitive therapy and research , 31 , 759-772. Moulding, R., Doron, G., Kyrios, M., & Nedeljkovic, M. (2008). Desire for control, sense of control and obsessive-compulsive checking: An extension to clinical samples. Journal of Anxiety Disorders , 22 (8), 1472-1479. Moulding, R., Kyrios, M., & Doron, G. (2007). Desire for control, sense of control and obsessive-compulsive symptoms. Cognitive therapy and research , 31 , 759-772. Moulding, R., Kyrios, M., & Doron, G. (2007). Obsessive-compulsive behaviours in specific situations: The relative influence of appraisals of control, responsibility and threat. Behaviour Research and Therapy , 45 (7), 1693-1702. Müller, M.J. (2012). Will it hurt less if I believe I can control it? Influence of actual and perceived control on perceived pain intensity in healthy male individuals: a randomized controlled study. J Behav Med 35 , 529–537. https://doi.org/10.1007/s10865-011-9382-0 Prolific. (2014). Prolific (Version 10.2023). London, UK. Available at https://www.prolific.com. Richards, H. J., Benson, V., Donnelly, N., & Hadwin, J. A. (2014). Exploring the function of selective attention and hypervigilance for threat in anxiety. Clinical psychology review , 34 (1), 1-13. Robins, R. W., Hendin, H. M., & Trzesniewski, K. H. (2001). Measuring global self-esteem: Construct validation of a single-item measure and the Rosenberg Self-Esteem Scale. Personality and social psychology bulletin , 27 (2), 151-161. Salkovskis, P. M. (1996). The cognitive approach to anxiety: Threat beliefs, safety-seeking behavior, and the special case of health anxiety and obsessions. In P. M. Salkovskis (Ed.), Frontiers of Cognitive Therapy (pp. 48–74). Guilford Press. Seligman, M. E. (1972). Learned helplessness. Annual review of medicine , 23 (1), 407-412. Shafran, R., & Mansell, W. (2001). Perfectionism and psychopathology: A review of research and treatment. Clinical psychology review , 21 (6), 879-906. Skinner, E. A. (1996). A guide to constructs of control. Journal of Personality and Social Psychology , 71 (3), 549. Soenens, B., Vansteenkiste, M., Luyten, P., Duriez, B., & Goossens, L. (2005). Maladaptive perfectionistic self-representations: The mediational link between psychological control and adjustment. Personality and individual differences , 38 (2), 487-498. Spitzer, R. L., Kroenke, K., Williams, J. B. W., & Löwe, B. (2006). A brief measure for assessing generalized anxiety disorder: The GAD-7. Archives of Internal Medicine, 166 (10), 1092–1097. DOI: 10.1001/archinte.166.10.1092 StataCorp. (2023). Stata Statistical Software: Release 19. College Station, TX: StataCorp LLC. Urben, S., Ochoa Williams, A., Ben Jemia, C., Rosselet Amoussou, J., Machado Lazaro, S., Giovannini, J., … & Mürner-Lavanchy, I. (2024). Understanding irritability through the lens of self-regulatory control processes in children and adolescents: a systematic review. European child & adolescent psychiatry , 1-13. Wells, A. (1995). Meta-cognition and worry: A cognitive model of generalized anxiety disorder. Behavioural and cognitive psychotherapy , 23 (3), 301-320. Variables Below GAD-7 Cutoff ( n = 897) At or Above GAD-7 Cutoff ( n = 428) Total ( N = 1,325) Sociodemographic Characteristics Age: Mean( SD ) 37.81(16.69) 31.91(13.51) 35.90(15.96) Gender Identity: % Woman 50.17% 64.49% 54.79% Man 49.28% 35.05% 44.68% Non-Binary 0.56% 0.47% 0.53% Race/Ethnicity: % Asian 15.16% 12.62% 14.34% Black/African American 10.70% 12.62% 11.32% Hispanic/Latino 8.70% 11.92% 9.74% White 53.29% 48.13% 51.62% Other/Not Listed 12.15% 14.71% 12.98% Mental Health Concerns: Mean(SD) Perceived Stress 2.13(0.84) 3.23(0.72) 2.48(0.95) Anxiety Symptoms 3.69(2.81) 14.79(3.41) 7.28(6.00) Depressive Symptoms 11.60(9.52) 27.43(13.52) 16.72(13.24) Control Beliefs: Mean(SD) Perceived Control 5.04(1.00) 4.12(0.95) 4.75(1.07) Desired Control 4.90(0.68) 4.73(0.65) 4.85(0.67) Control Discrepancy 4.85(1.55) 4.39(1.80) 4.70(1.65) Note: SD = standard deviation. Figure 1. Distribution of global control discrepancy scores by GAD cutoff categorization. Figure 2. Receiver Operating Characteristic (ROC) Curve for Predicting GAD Risk. math_shortcuts Figure 3. Logistic regression model illustrating the probability of scoring at or above the GAD-7 cutoff point based on self-reported magnitude of control discrepancy at different levels of perceived and desired control. Note: Model is adjusted for perceived control, desired control, perceived control x desired control, depressive symptoms, gender identification, and age. 95% CI Pairwise Comparisons ∆ SE Z p Lower Upper “Extremely less than I desire” vs. “Extremely more than I desire” -0.20 0.04 -4.62 < .001 -0.29 -0.11 “Much less than I desire” vs. “Extremely more than I desire” -0.23 0.04 -5.92 < .001 -0.31 -0.16 “Somewhat less than I desire” vs. “Extremely more than I desire” -0.25 0.04 -6.18 < .001 -0.32 -0.17 “Slightly less than I desire” vs. “Extremely more than I desire” -0.25 0.04 -5.96 < .001 -0.33 -0.16 \raggedright “As much as I desire” vs. “Extremely more than I desire” \raggedright -0.23 \raggedright 0.04 \raggedright -5.64 \raggedright < .001 \raggedright -0.31 \raggedright -0.15 “Slightly more than I desire” vs. “Extremely more than I desire” -0.20 0.04 -5.34 < .001 -0.27 -0.13 “Somewhat more than I desire” vs. “Extremely more than I desire” -0.15 0.03 -5.11 < .001 -0.21 -0.09 “Much more than I desire” vs. “Extremely more than I desire” -0.08 0.02 -4.98 < .001 -0.12 -0.05 Note: ∆ = delta; SE = standard error; z = z-statistic; CI = confidence interval. Model is adjusted for perceived control, desired control, depressive symptoms, age, and gender identification. Figure 4. Ordered logistic regression model depicting the probability of severe GAD symptom severity category levels across standardized control discrepancy scores at different levels of perceived and desired control. Note: Model is adjusted for perceived stress, depressive symptoms, age, and gender identification. Predictor Wald χ 2 p > χ 2 Log Likelihood R 2 ∆ R 2 OR(SE) z p [95% CI] Block 1: Known Control Beliefs 185.84 < .001 -608.37 .17 — Intercept .35(.03) -14.27 < .001 [0.31, 0.41] Perceived control .31(03) -13.05 < .001 [0.26, 0.37] Desired control 1.17(.09) 2.11 .035 [1.01, 1.36] Block 2: Interaction 176.13 < .001 -607.19 .18 .01 Intercept .36(.03) -13.21 < .001 [0.31, 0.42] Perceived control .31(.03) -13.06 < .001 [0.26, 0.37] Desired control 1.13(.09) 1.49 .137 [0.96, 1.32] Perceived control x desired control .89(.07) -1.51 .130 [0.76, 1.04] Block 3: Known Risk Factors 383.43 < .001 -332.42 .55 .37 Intercept .26(.04) -9.22 < .001 [0.19, 0.34] Perceived control .87(.13) -0.91 .364 [0.65, 1.17] Desired control 1.28(.15) 2.08 .037 [1.02, 1.60] Perceived control x desired control .68(.07) -3.68 < .001 [0.56, 0.84] Depressive Symptoms 6.33(1.10) 10.66 < .001 [4.51, 8.89] Perceived Stress 3.81(.69) 7.39 < .001 [2.67, 5.43] Block 4: Novel Control Belief 353.30 < .001 -315.14 .57 .02 Intercept .22(.04) -8.84 < .001 [0.16, 0.31] Perceived control .74(.12) -1.81 .070 [0.54, 1.03] Desired control 1.29(.16) 2.01 .045 [1.01, 1.64] Perceived control x desired control .65(.07) -3.99 < .001 [0.52, 0.80] Depressive Symptoms 7.50(1.33) 11.36 < .001 [5.30, 10.62] Perceived Stress 4.40(.85) 7.68 < .001 [3.02, 6.42] Control discrepancy (linear) 1.02(.22) 0.10 .919 [0.67, 1.56] Control discrepancy (quadratic) 1.02(.09) 0.23 .819 [0.86, 1.21] Control discrepancy (cubic) 1.18(.07) 2.79 .005 [1.05, 1.33] Note: χ 2 = chi-square, ∆R 2 = change in R 2 , OR = observed odds ratio, SE = bootstrapped standard error, z = z-statistic, CI=confidence interval. Blocks 3 and 4 are adjusted for age and gender identification. Supplementary Figure S1. Proportions of accuracy (solid line), sensitivity (dashed line), and specificity (dotted line) are plotted across probability thresholds ranging from 0.20 to 0.80. Information & Authors Information Version history V1 Version 1 23 June 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords anxiety appraisals cognitive orientation Authors Affiliations Christopher J. Davis 0009-0005-2859-5354 [email protected] Cornell University Department of Psychology View all articles by this author Sydnie R. Spearman The Chicago School View all articles by this author Metrics & Citations Metrics Article Usage 475 views 166 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Christopher J. Davis, Sydnie R. Spearman. Control Discrepancy and Anxiety Symptomatology: A Distinct Metacognitive Vulnerability Beyond Traditional Control Constructs. Authorea . 23 June 2025. DOI: https://doi.org/10.22541/au.175067477.74668572/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); Cited by Christopher J. Davis, Sydnie R. Spearman, Anthony L. Burrow, Perceived, Desired, and Discrepant Control: Initial Cross-Sectional Evidence for a Metacognitive Calibration Framework for Depression Risk, Cognitive Therapy and Research, (2026). https://doi.org/10.1007/s10608-026-10725-2 Crossref Loading... View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.175067477.74668572/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9fe78df1c8521640',t:'MTc3OTI0MDM5OQ=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();