The Role of Perceived Control in the Stress–Anxiety Pathway: Exploring Acute Temporal Dynamics and Contextual Boundaries

Abstract

Perceived control has been identified as a key trait that buffers the stress–anxiety pathway. However, the extent to which perceived control exerts its protective effect across varying magnitudes of stressors remains understudied. Moreover, questions remain regarding when and how perceived control disrupts the stress–anxiety link. In the present study, we examine how perceived control influences subjective stress and state anxiety under various levels of stress exposure. Two hundred twenty-five adults (age: M = 26.4 years, SD = 10.8) were randomly assigned to either a low-stress or high-stress condition. Linear regressions, linear mixed-effects models, and moderated mediation path analyses were performed. Perceived control attenuated the effects of stress exposure on subjective stress and anxiety symptoms only in the high-stress condition. Path analyses revealed that perceived control indirectly influenced anxiety symptoms by reducing subjective-stress reactivity. Perceived control may buffer against the psychological impact of high-stress contexts by reducing subjective-stress reactivity to prevent downstream anxiety symptoms.

Keywords

perceived control subjective stress anxiety symptoms stress exposure open data open materials preregistered

The interplay between stress and anxiety has been widely studied in efforts to understand how individuals manage and respond to environmental demands. Central to this area is the role of perceived control, defined as the extent to which people believe in their capacity to influence their life circumstances and bring about desired outcomes (Skinner, 1996). Although perceived control has been consistently demonstrated to buffer the effects of stress–related outcomes (Barlow, 2004; Zhao et al., 2021; Zheng et al., 2020), recent clinical studies have highlighted its critical role in anxiety interventions (Gallagher, Bentley, & Barlow, 2014; Gallagher, Naragon-Gainey, & Brown, 2014; Wadsworth et al., 2019), reigniting interest in how perceived control exerts its influence. However, insights regarding how and when perceived control is impactful in the stress–anxiety pathway remain underexplored. Moreover, few studies have tested its contextual boundaries or indirect pathways in momentary contexts of stress. Guided by Lazarus and Folkman’s (1984) transactional model of stress and coping, in the present study, we examine how perceived control shapes the relationship between stress exposure at varying levels (i.e., high and low) and stress–related outcomes (i.e., subjective stress and state-anxiety symptoms), clarifying the temporal and contextual boundaries of perceived control as a protective factor in stress regulation. In addition, we examine the direct and indirect pathways regarding how and when perceived control exerts its protective influence in the stress–anxiety pathway. These insights may offer necessary refinement to existing theoretical models and intervention efforts at the clinical and population levels.

Stress and Anxiety: A Transactional Framework

The stress–anxiety pathway is widely recognized as bidirectional and interdependent, and stress is considered an integral precipitating factor in the development and maintenance of anxiety symptoms. Indeed, it is clear that one’s experience of stress (e.g., subjective, physiological) may affect one’s risk for future anxiety-related outcomes (Boi & Llera, 2023; Daviu et al., 2019; Qi et al., 2016). For instance, Tolin et al. (2021) found that individuals with anxiety disorders exhibit unique stress–response patterns (e.g., individuals with generalized anxiety disorders demonstrated reduced high-frequency heart rate variability compared with their counterparts). Moreover, chronic stress has been shown to sensitize neural circuits in the amygdala and prefrontal cortex, which may amplify threat perceptions and increase the likelihood of developing anxiety symptoms (Delgado et al., 2008). Researchers have also clarified the role of specific individual differences (e.g., intolerance of uncertainty or feelings of helplessness) in one’s subjective experience of stress and downstream anxiety-related outcomes (e.g., an individual with high intolerance of uncertainty may experience more intense threat-related appraisals and impaired coping when engaging with a task that may yield unpredictable outcomes; Sahib et al., 2023).

Lazarus and Folkman’s (1984) transactional model of stress and coping (commonly referred to as the “transactional-stress theory”) provides a foundational framework for understanding the role of subjective appraisals in the stress–anxiety pathway. According to the transactional-stress theory, stress emerges from person–environment interactions, in which the perception of a threat (i.e., primary appraisal) and evaluation of coping resources (i.e., secondary appraisal) heavily influence one’s stress response. In this model, Lazarus and Folkman maintained that stress appraisals precede anxiety such that anxiety emerges downstream of individuals’ evaluations of situational demands relative to perceived coping resources. Fittingly, when individuals appraise a stressor as overwhelming or beyond their capacity to effectively cope, their psychological and physiological stress intensifies, thereby increasing vulnerability to anxiety (Biggs et al., 2017; Lazarus & Folkman, 1984). Recent empirical evidence supports key predictions of this theory, particularly concerning the role of psychological factors (e.g., self-efficacy, rejection sensitivity) in shaping stress and anxiety responses (Gao et al., 2017; Jerusalem & Schwarzer, 2014). Although perceived control has been widely examined in this context, important questions remain about its specificity, timing, and limits as a protective factor.

Perceived Control as a Protective Factor

Perceived control—broadly defined as individuals’ belief in their ability to influence their life outcomes—has regularly demonstrated its buffering role against the adverse effects of stress and anxiety-related outcomes (Lachman & Weaver, 1998; Skinner, 1996). High perceived control has been associated with reduced physiological reactivity to stressors (Wen & Sin, 2022; Zilioli et al., 2017) and lower risk of developing anxiety disorders (Chorpita & Barlow, 1998). Note that perceived control is thought to influence stress and anxiety levels through both cognitive appraisals and coping processes. For instance, previous research has demonstrated that one’s sense of agency (a core component of perceived control in many conceptualizations; see Lachman & Weaver, 1998) enhances one’s confidence in one’s ability to manage stressors, which may decrease the perceived intensity of a stressor (Bandura, 2015; Zajacova et al., 2005). In addition, individuals with high perceived control are more likely to engage in adaptive coping strategies (e.g., problem-focused coping), which are associated with better psychological outcomes compared with maladaptive strategies (e.g., avoidant or emotion-focused coping; Dijkstra & Homan, 2016). Although high perceived control has been evidenced to be beneficial to stress and anxiety outcomes, previous research has also demonstrated the impact of low perceived control, and some researchers have identified it as a vulnerability to anxiety disorders (Gallagher, Bentley, & Barlow, 2014; Liang et al., 2025; Wadsworth & Hayes-Skelton, 2020).

Much of the research on the interplay between perceived control, stress, and anxiety-related outcomes has operationalized perceived control as either an individual difference (e.g., a relatively stable, enduring personality trait) or a contextual affordance (e.g., an appraisal of control based on environmental or situational conditions, such as access to resources that may represent objective control). Both approaches have demonstrated the role of perceived control in influencing physiological and psychological reactions to stress (Bhanji & Delgado, 2014; Bollini et al., 2004; Gilchrist et al., 2015; Halford et al., 2012; Müller, 2012). For example, Bollini et al. (2004) found that participants with higher levels of trait perceived control had a reduced cortisol response when placed in a condition that afforded objective control over the study stressor compared with participants who were not. Similar effects have been found when considering perceived control only as a contextual affordance. Gilchrist et al. (2015) found that participants who could choose when to take a break during a surgery video known to elicit vasovagal responses reported fewer vasovagal symptoms than participants who could not select their break time. These findings, particularly those regarding perceived control as an individual difference, have recently extended to clinical populations. In a sample of partially hospitalized individuals receiving psychotherapy (e.g., group and individual) and pharmacological intervention, Wadsworth et al. (2019) found perceived control increased during a brief intensive treatment, which corresponded with decreased anxiety symptoms. Likewise, Gallagher, Bentley, and Barlow (2014) found perceived control to be a mechanism of change in cognitive-behavior-therapy outcomes for anxiety disorders such that increases in perceived control led to decreased anxiety symptoms over the course of treatment.

Gaps in the Literature: Temporal Dynamics and Contextual Boundaries

Although the benefits of perceived control on stress and anxiety-related outcomes have been well established, two critical gaps in the literature remain. First, the temporal dynamics of perceived control’s protective effects are underexplored. Specifically, it is unclear whether perceived control exerts its influence primarily during the initial stages of stress appraisal or if it continues to affect anxiety symptoms over time. Given the causal role of perceived control in anxiety-related outcomes, further clarity is needed regarding how perceived control exerts its influence. Clarifying this dynamic may inform theoretical frameworks and treatment approaches.

Second, the contextual boundaries of perceived control’s effect on the stress–anxiety pathway are not well understood. It is relatively unknown whether the benefits of perceived control are consistent across low-stress and high-stress environments or whether they emerge or become more pronounced under specific levels of subjective stress. This is particularly noteworthy because emerging research has highlighted the differential impact of perceived control based on context or life circumstances (e.g., socially disadvantaged populations) such that perceived control is more beneficial in adverse contexts. For instance, Turiano et al. (2014) found that the predictive role of perceived control on mortality rate was evident only among individuals low in years of education such that individuals endorsing high perceived control and low levels of educational attainment had a lower 14-year mortality risk than individuals endorsing low perceived control and low levels of educational attainment. At greater levels of educational attainment, perceived control was not associated with mortality risk. Likewise, there has been evidence put forth regarding the role of perceived control in downstream implications on mental-health symptoms, specifically, depressive and anxiety symptoms. Davis and Burrow (2024) found that perceived control attenuated the relationship between adverse childhood experiences (ACEs) and internalizing symptoms (i.e., depressive and anxiety symptoms) such that individuals reporting high perceived control and high levels of ACEs reported significantly fewer internalizing symptoms than individuals with the same level of ACEs but either moderate or low perceived control. Conversely, they found no significant difference in internalizing symptoms at any level of perceived control (i.e., low, moderate, or high) for individuals reporting low levels or no ACEs. These examples highlight the potential for perceived control to emerge as a protective factor in high-stress or adverse contexts compared with low-stress or favorable environments. Thus, it seems possible (and perhaps likely) that the benefits of perceived control on the stress–anxiety pathway are evident only in contexts of increased stress but less impactful when contextual stress is low.

Present Study

Grounded in Lazarus and Folkman’s (1984) transactional-stress theory, in the present study, we examine the role of perceived control in the stress–anxiety pathway. Specifically, we explore the moderating role of perceived control in the relationship between stress exposure (i.e., high- and low-stress conditions) and subjective stress and state-anxiety symptoms. To address meaningful gaps in the literature, we also examine whether perceived control exerts its protective effects similarly across levels of stress exposure and explore its specific influence on anxiety symptoms and the way it influences anxiety symptoms. Accordingly, in the present study, we propose the following hypotheses:

Hypothesis 1 (H1): Perceived control attenuates the positive relationship between stress exposure and Time 2 (T2) subjective stress. In addition, perceived control has no influence on T2 subjective stress in the low-stress condition but significant influence in the high-stress condition.

Hypothesis 2 (H2): Perceived control attenuates the positive relationship between stress exposure and posttest state-anxiety symptoms. In addition, perceived control has no association with posttest state-anxiety symptoms in the low-stress condition but a significant association in the high-stress condition.

(Exploratory) Hypothesis 3 (H3): Perceived control moderates the path from stress exposure to subjective-stress reactivity as well as the path from subjective-stress reactivity to posttest state-anxiety symptoms.¹

Method

Participants

The sample comprised 225 adults (age: M = 27.6 years, SD = 11.2, range = 18–65) recruited through Prolific—an online platform designed to connect researchers with participants for academic and scientific studies (www.prolific.com).² Initial inclusion criteria were as follows: (a) having at least a 95% approval rating across at least 100 total submissions on Prolific; (b) being at least 18 years of age, (c) fluent in English, and (d) located in the United States; and (e) answering “Yes” to the question “Would you take part in a study where you are intentionally given inaccurate information about other participants and the study? You would be debriefed after the study.”³ Exclusion criteria were (a) obtaining a ReCaptcha score < .7 (Griffin et al., 2022), (b) completing the cognitive-testing blocks in at least 3 SD below the median completion time, and (c) failing to fully completing each measure of interest.⁴ After enforcing exclusion criteria, our final analytic sample comprised 214 adults (age: M = 26.4 years, SD = 10.8, range = 18–65). Table 1 reports detailed demographic characteristics and descriptive statistics of participants included in the final analytic sample by condition assignment. Demographic characteristics were similar between study conditions. Perceived control, pretest subjective stress, confidence in cognitive abilities, trait-anxiety symptoms, and age were balanced between conditions. As expected, individuals in the high-stress condition reported higher posttest state-anxiety symptoms and posttest subjective stress, on average, than individuals in the low-stress condition.

Table 1.

Demographic Characteristics and Descriptive Statistics

Variable	Low-stress condition(n = 105)	High-stress condition(n = 109)	Total(N = 214)
Sociodemographic characteristics
Age, M (SD)	26.08 (10.60)	26.66 (10.96)	26.37 (10.76)
Sex, %
Female	66.46%	68.13%	67.29%
Male	32.30%	30.63%	31.46%
Neither female nor male	1.24%	0.63%	0.93%
Prefer not to mention	0.00%	0.63%	0.31%
Annual household income, %
Less than $50,000	27.04%	31.88%	29.47%
$50,000–$74,999	18.24%	15.00%	16.61%
$75,000–$99,999	11.32%	10.63%	10.97%
$100,000–$149,999	16.35%	14.38%	15.36%
More than $149,999	27.04%	28.13%	27.59%
Race/ethnicity, %
Asian	12.22%	21.90%	17.23%
Black/African American	16.48%	16.03%	16.25%
Hispanic/Latino	10.19%	12.07%	11.16%
White	54.63%	43.97%	49.11%
Other/not listed	6.48%	6.03%	6.25%
Control beliefs, M (SD)
Perceived control	4.61 (1.16)	4.85 (1.09)	4.73 (1.12)
Confidence in cognitive abilities	7.25 (1.86)	7.58 (1.73)	7.42 (1.80)
Subjective stress, M (SD)
Pretest subjective stress	49.06 (28.99)	47.93 (27.45)	48.47 (28.14)
Posttest subjective stress	22.48 (23.38)	58.08 (29.15)	40.92 (31.91)
Changes in subjective stress	−26.57 (32.59)	10.15 (30.15)	−7.56 (36.28)
Anxiety symptoms, M (SD)
State-anxiety symptoms	1.37 (0.48)	1.71 (0.68)	1.55 (0.62)
Trait-anxiety symptoms	2.42 (0.88)	2.25 (0.83)	2.34 (0.86)

Power analysis

Two complementary power analyses were conducted after preregistration was time-stamped and before any primary analyses were conducted to assess the adequacy of the available sample. First, power for the focal interaction terms in our regression models was estimated using G*Power 3.1 (Faul et al., 2009). For H1, assuming a medium effect size ( f² = 0.15), α = .05, six predictors, and 90% power, the required sample was N = 123. For H2, assuming a small to medium effect size (f² = 0.10), α = .05, six predictors, and 90% power, the required sample was N = 181. Second, for the full structural equation model (SEM), a Monte Carlo simulation-based power analysis were conducted (1,000 replications per sample size, Ns = 100–1,000 in increments of 50). Results indicated that five of the six paths achieved ≥ .80 power by N = 150 and that all six reached ≥ .90 power by N = 200.

Measures

Perceived control

Perceived control was assessed using the Sense of Control Scale (SOCS; Lachman & Weaver, 1998). The SOCS is a 12-item self-report measure scored on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree) in which respondents are asked to report the extent to which they agree with each statement. The SOCS includes two subscales: Personal Mastery and Perceived Constraints. An example item from the Personal Mastery subscale is “When I really want to do something, I usually find a way to succeed at it.” An example item from the Perceived Constraints subscale is “I have little control over the things that happen to me.” The SOCS is calculated by first reverse-scoring each item on the Perceived Constraints subscale (so that higher scores indicate a lower sense of perceived constraints), summing each item, and then dividing by the number of items. For interpreting perceived control, higher scores indicate a greater sense of control over one’s life. In the present study, the SOCS demonstrated excellent internal consistency (α = .91).

Subjective stress (pretest and posttest)

The Visual Analogue Scale for Stress (VAS-S; Lesage et al., 2012) was used to assess subjective stress both before and after manipulation. The VAS-S is a single-item self-report measure in which respondents are asked to report the extent to which they presently feel stressed (e.g., “Presently, how stressed do you feel?”). The VAS-S is scored on an analogue scale ranging from 0 to 100. Previous work has shown the VAS-S to be at least as discriminating as commonly used stress questionnaires (e.g., Perceived Stress Scale) in observing differences in subjective stress between groups (Barré et al., 2017; Lesage et al., 2012). Furthermore, previous studies have utilized the VAS-S in experimental designs similar to the present study to measure pre-to-post differences in subjective stress (Binfet et al., 2018; Fuentes-García et al., 2019; Ilioudi et al., 2023).

Subjective-stress reactivity

Subjective-stress reactivity was calculated through estimating residualized change scores by regressing T2 subjective stress on Time 1 (T1) subjective stress. This score indicates the extent to which individuals’ subjective stress at T2 deviates from what would be expected based on their baseline (T1) subjective stress.

Posttest state-anxiety symptoms and baseline trait-anxiety symptoms

Posttest state-anxiety symptoms and trait-anxiety symptoms were assessed through the State-Trait Anxiety Inventory–10 (STAI-10; Spielberger et al., 1983). The STAI-10 comprises two subscales: State Anxiety and Trait Anxiety. Each subscale contains five items and is scored on a 4-point Likert scale ranging from 1 (not at all) to 4 (very much). Example items from the State Anxiety subscale are “I feel nervous” and “I feel jittery.” Example items from the Trait Anxiety subscale are “I feel that difficulties are piling up so high that I cannot overcome them” and “I get in a state of tension or turmoil as I think over my recent concerns and interests.” In the present study, the State Anxiety subscale demonstrated good internal consistency (α = .81), and the Trait Anxiety subscale demonstrated excellent internal consistency (α = .90).

Confidence in cognitive abilities

A single item assessing participants’ confidence in their cognitive abilities was developed and included as a covariate. The item reads, “Please rate the extent to which you agree with the following statement: I am confident in my cognitive abilities.” This item is scored on a 10-point Likert scale ranging from 1 (strongly disagree) to 10 (strongly agree). Previous studies have demonstrated the utility of single-item measures compared with more exhaustive scales aimed at measuring individual differences (Postmes et al., 2013; Robins et al., 2001).

Procedure

Figure 1 presents a flowchart illustrating the study procedures. Before enrolling in the present study, participants were made aware of the study by a posting on Prolific’s participant board titled “Answer a survey about your personality and complete a few cognitive activities” with the following caption:

The purpose of this study is to better understand how differences in personality may impact the testing experience of our users. You will take a series of brief surveys and test out a few of our cognitive activities. This should take about 20 minutes.

Fig. 1.

Flowchart of study procedures.

Informed consent, pretest measures, and randomization

Participants first read and signed a consent form approved by the Institutional Review Board (IRB). The consent form provided deceptive information about study procedures (for informed consent used in the present study, see study repository: osf.io/zn6sb/). Participants then completed pretest assessments. Next, participants were randomly assigned to either the control (low stress) or experimental (high stress) condition through Qualtrics’s (www.qualtrics.com) randomizer function.⁵

Experimental manipulation

To manipulate subjective stress (e.g., increase subjective stress in the high-stress condition but not the low-stress condition), the online, autonomous version of the Trier Social Stress Test (Gunnar et al., 2021) was adapted for this study. This approach has been used in previous research and has demonstrated efficacy in inducing an increased subjective-stress response to the experimental condition (see Davis et al., 2024). In this approach, both the low-stress and high-stress conditions were assigned with three blocks of cognitive tests, and participants were given predetermined deceptive feedback on their performance after each test. Participants in the low-stress condition were assigned markedly easier questions (e.g., “Charles shoots an arrow 50 yards. Sally shoots his arrow three times as far. How far did Sally’s arrow go?”) and no feedback on their performance at the end of each testing block (e.g., “You’ve just completed the 1st block of questions. You will now be directed to the 2nd block of questions”). However, participants in the high-stress condition were assigned more challenging questions (e.g., “Which two numbers should replace the question marks: 165-162-3, 61-54-7, 30-18-12, ?-?-18”) and given negative feedback on their performance at the end of each testing block (e.g., “You scored higher than only 9 percent of users. You scored lower than 91 percent of users.”). Test questions in each block for the low-stress and high-stress conditions are available in our study repository. In addition, participants in the high-stress condition were given a time limit of 3 min per testing block and received the message, “WARNING: Your performance is being monitored,” whereas participants in the low-stress condition did not.⁶

Posttest assessment and debriefing

All participants completed posttest assessments directly after completing the final testing block. Finally, participants were informed of the true purpose of the experiment and the exact nature of the deceptions used throughout the study. The full debriefing page is available in our study repository.

Transparency and Openness

Preregistration

The present study was preregistered through OSF (osf.io/hcbgs/) before data analysis. We used data from a larger series of experiments conducted by our research group that were also preregistered through OSF (osf.io/a4c9f) before data collection.

Data, materials, code, and online resources

All study-related material, including survey material, analytic code, data output for each model, and raw data supporting the findings of this study, are available through the study repository on OSF (osf.io/zn6sb/).

Reporting

We report how we determined our sample size, all data exclusions, all manipulations, and all measures used in the present study.

Ethical approval

All study procedures received institutional approval from the Cornell University IRB (IRB No. 0147235) and was conducted in accordance with provisions of the Declaration of Helsinki.

Data Analytic Strategy

All analyses were conducted in Stata BE/19.5 (StataCorp, 2023). Before hypothesis testing, Little’s missing completely at random (MCAR) test was conducted to assess whether the pattern of missing data conformed to the assumption of MCAR.

Manipulation check

To assess the efficacy of the experimental manipulation, we fit a linear mixed-effects model with both fixed and random effects to analyze changes in subjective stress. Fixed effects included time (pre/post), condition (high vs. low stress), their interaction, baseline subjective stress, and trait-anxiety symptoms. Random intercepts and slopes for time were included for each participant to account for individual differences in baseline stress levels and change over time. Robust standard errors were clustered by participant.⁷

H1 and H2: perceived control attenuates the relationship between stress exposure and T2 subjective stress (H1) and posttest state-anxiety symptoms (H2)

Before hypothesis testing, continuous predictor variables were standardized to enhance coefficient comparability and mitigate potential multicollinearity concerns (Harrell, 2001). A Bonferroni correction was applied to address multiple hypothesis testing, resulting in an adjusted significance threshold of p = .025. Hypotheses 1 and 2 were examined through an ordinary least squares multiple-linear-regression approach, and bootstrapped standard errors (5,000 replications) were estimated. Models testing H1 and H2 included stress exposure (reference = low stress), perceived control, and their interaction; covariates were confidence in cognitive abilities, age, and trait-anxiety symptoms.⁸ Cook’s distance was used to identify potential outliers. Data points reporting a Cook’s distance value greater than $\frac{4}{n}$ were excluded from future model iterations. Finally, we conducted a simple marginal-effects analysis of the best-performing H1 and H2 models to examine and illustrate within-conditions and between-conditions effects of perceived control.

H3: perceived control is a moderates the mediated pathway

H3 was examined through a moderated mediation path analysis across two models. Model 3A included direct paths from (Path 1) stress exposure (reference = low-stress condition) to subjective-stress reactivity and (Path 2) posttest state anxiety, (Path 3) subjective-stress reactivity to posttest state-anxiety symptoms, and (Path 4) trait-anxiety symptoms to posttest state-anxiety symptoms. In addition, we included perceived control as a moderated mediator in Paths 1 and 3. Model 3B included Model 3A paths but excluded the direct path from stress exposure to posttest state anxiety (Path 2) and perceived control as a moderated mediator in the path from subjective-stress reactivity and posttest state-anxiety symptoms (Path 4). A likelihood-ratio test was conducted, and Akaike information criterion (AIC) and Bayesian information criterion (BIC) values between Model 3A and Model 3B were compared to determine the better-performing model. Model fit was assessed using established thresholds (see Hu & Bentler, 1999) based on several fit indices: comparative fit index (CFI) > .90, Tucker-Lewis index (TLI) > .90, root mean square error of approximation (RMSEA) < .05, and standardized root mean squared residual (SRMR) < .05. All adjustments to improve model fit were guided by modification indices and theoretical considerations. Before fitting the SEMs, trait-anxiety symptoms, posttest state-anxiety symptoms, and perceived control were modeled as latent variables using their respective multiitem indicators to obtain bias-corrected estimates that account for measurement error. In contrast, subjective-stress reactivity was assessed with a residualized change score comprising repeated single items, and stress exposure was operationalized as a categorical experimental condition; thus, both variables were modeled as observed because they could not be specified as latent constructs.

Results

Preliminary analyses

Final analytic sample

Missing data were found to range between 0.00% and 4.44% for any single measure. Little’s MCAR test suggested that data were MCAR, χ²(6) = 3.36, p = .763; thus, listwise deletion was deemed appropriate to handle missing data.

Manipulation check

To examine the efficacy of our stress manipulation, a linear mixed-effects model (LMM) with both random and fixed effects was estimated, χ²(5) = 1,766.64, p < .001, log pseudolikelihood = −1,870.47. After controlling for baseline subjective stress and trait anxiety, we found that participants in the high-stress condition reported significantly greater subjective stress at T2 than participants in the low-stress condition (b = 36.58, SE = 4.29, z = 8.53, p < .001). There were no between-conditions differences in T1 subjective stress, χ²(1) = 0.06, p = .801, indicating successful randomization. Because T1 subjective stress was assessed after participants learned they would complete an IQ-style task, baseline subjective stress may have been elevated in both groups because of anticipatory stress, as shown in previous studies (Dickerson & Kemeny, 2004; Gruenewald et al., 2004). This may have produced decreases in the low-stress condition but increases in the high-stress condition as the task demands diverged. Nevertheless, these findings suggest that the stress manipulation was efficacious in both increasing subjective stress in the high-stress condition and facilitating overall between-conditions differences in subjective stress (see Fig. 2).

Fig. 2.

Changes in subjective stress as a function of time point and condition assignment.

Correlations and unadjusted relationships

A full correlation matrix with associated p values is provided in the online supplementary materials (osf.io/zn6sb/). Perceived control was inversely correlated with poststressor stress in the high-stress condition (r = −.36, p < .001) but not in the low-stress condition (r = −.10, p = .32). A similar pattern emerged for state anxiety. Perceived control was negatively correlated with state anxiety in both conditions, although more strongly under high stress (r = −.36, p < .001) than low stress (r = −.23, p = .019). In addition, consistent with theoretical distinctions between stress reactivity and anxiety responses, T2 subjective stress and state-anxiety symptoms were strongly and positively correlated in both the low-stress condition (r = .52, p < .001) and the high-stress condition (r = .54, p < .001), indicating substantial covariation but not redundancy (see Tabachnick et al., 2019).

Finally, to examine unadjusted predictors of T2 subjective stress and state-anxiety symptoms simultaneously, we estimated a seemingly unrelated regression model with perceived control, stress exposure (reference = low-stress condition), and their interaction predicting both outcomes. Stress exposure significantly predicted both T2 subjective stress (b = 37.72, SE = 3.50, p < .001) and state-anxiety symptoms (b = 2.00, SE = 0.38, p < .001). The Perceived Control × Stress Exposure interaction was significant for both T2 subjective stress (b = −9.48, SE = 3.76, p = .012) and state-anxiety symptoms (b = −0.83, SE = 0.41, p = .042), indicating that higher perceived control attenuated subjective stress and state-anxiety symptoms in the high-stress condition. A cross-equation Wald test indicated that the Perceived Control × Stress Exposure interaction was significantly stronger for T2 subjective stress than for state-anxiety symptoms, χ²(1) = 5.83, p = .016.

H1: Perceived control moderates the relationship between stress exposure and T2 subjective stress

Direct effects

In Model 1A, Wald χ²(5) = 271.46, R² = .49, Normalized Root Mean Square Error (NRMSE) = .123, we found a positive direct relationship between stress exposure (reference = low-stress condition) and T2 subjective stress (b = 38.34, SE = 3.19, z = 12.01, p < .001) but no direct relationship between perceived control and T2 subjective stress, although it trended toward significance (β = −3.25 SE = 1.79, z = −1.81, p = .070).⁹

Moderation effect

Consistent with our hypothesis, Model 1B, Wald χ²(6) = 313.99, R² = .51, NRMSE = .122, found perceived control attenuated the relationship between stress exposure (reference = low stress) and T2 subjective stress (b = −9.44, SE = 3.16, z = −2.99, p = .003) such that increased perceived control weakened the positive relationship between stress exposure and changes in subjective stress. The left graph in Figure 3 illustrates this finding.

Fig. 3.

Marginal effects results from (left) Model 1B and (right) Model 2B.

Between-conditions differences

We also found between-conditions differences in T2 subjective stress at each level of perceived control (low: χ² = 131.10, p < .001; moderate: χ² = 149.72, p < .001; high: χ² = 43.50, p < .001). Furthermore, we found diminished between-conditions differences in T2 subjective stress at increasing levels of perceived control. Specifically, when comparing participants in the low-stress condition with participants in the high-stress condition, we found that participants endorsing low perceived control demonstrated the greatest between-conditions difference in T2 subjective stress ( $\frac{d y}{d x}$ = 47.90, SE = 4.18) compared with participants endorsing either moderate perceived control ( $\frac{d y}{d x}$ = 38.94, SE = 3.18) or high perceived control ( $\frac{d y}{d x}$ = 29.99, SE = 4.55).

Within-conditions differences

An examination of the within-conditions effect of perceived control on T2 subjective stress offered further unique insights into the nature of this relationship. We found that perceived control did not affect T2 subjective stress for participants in the low-stress condition ( $\frac{d y}{d x}$ = 1.38, SE = 2.33, z = .59, p = .555) but had a negative effect on T2 subjective stress in the high-stress condition ( $\frac{d y}{d x}$ −8.06, SE = 2.35, z = −3.44, p = .001). These findings offer support to our hypothesis that the effect of perceived control on subjective stress is limited to contexts of heightened stress, and no evidence of an effect within contexts of lower stress was found.

H2: perceived control moderates the relationship between stress exposure and posttest state-anxiety symptoms

Direct effects

Model 2A, Wald χ²(5) = 87.29, R² = .30, NRMSE = .201, found a positive relationship between stress exposure (reference = low-stress condition) and posttest state-anxiety symptoms (b = 1.40, SE = 0.29, z = 4.82, p < .001) but no relationship between perceived control and state-anxiety symptoms (β = −0.29, SE = 0.16, z = −1.79, p = .074), although it trended toward significance.¹⁰

Moderation effect

Consistent with our hypothesis, Model 2B, Wald χ²(6) = 91.98, R² = .36, NRMSE = .193, found perceived control to attenuate the relationship between stress exposure (reference = low-stress condition) and posttest state-anxiety symptoms (β = −0.81, SE = 0.29, z = −2.81, p = .005) such that perceived control weakened the positive relationship between stress exposure and posttest state-anxiety symptoms. The right graph in Figure 3 illustrates this finding.

Between-conditions differences

Similar to our finding from H1, we found between-conditions differences in posttest state-anxiety symptoms at each level of perceived control (low: χ² = 31.10, p < .001; moderate: χ² = 39.03, p < .001; high: χ² = 12.08, p = .001). Furthermore, we found diminished between-conditions differences in posttest state-anxiety symptoms at increasing levels of perceived control such that participants reporting low perceived control reported the greatest between-conditions difference in posttest state-anxiety symptoms ( $\frac{d y}{d x}$ = 2.50, SE = 0.45) compared with participants endorsing either moderate perceived control ( $\frac{d y}{d x}$ = 1.78, SE = 0.29) or high perceived control ( $\frac{d y}{d x}$ = 1.06, SE = 0.30).

Within-conditions differences

In examining the effect of perceived control within both conditions, we found no significant effect of perceived control on posttest state-anxiety symptoms for participants in the low-stress condition ( $\frac{d y}{d x}$ = 0.05, SE = 0.15, z = 0.31, p = .758) but a negative effect on posttest state-anxiety symptoms ( $\frac{d y}{d x}$ −0.77, SE = 0.26, z = −2.92, p = .004) for participants in the high-stress condition. These findings offer further support to our hypothesis that the effect of perceived control on state-anxiety symptoms is evident within contexts of heightened stress but not contexts of lower stress.

H3: perceived control moderates the mediated path from stress exposure to subjective-stress reactivity but not state-anxiety symptoms

A moderated mediation path analysis was conducted to offer more granular insights into the temporal nature of the influence of perceived control on subjective stress and state-anxiety symptoms throughout the experiment. Accordingly, we examined the role of perceived control as a moderated mediator in the relationships between stress exposure, subjective-stress reactivity, and posttest state-anxiety symptoms while simultaneously modeling the influence of trait-anxiety symptoms. See Figure 4 for the paths and model-fit indices, slope coefficients, bootstrapped standard errors, and significance levels of the retained Model 3B.

Fig. 4.

Moderated mediation path analysis of Model 3B. Trait anxiety, posttest state anxiety, and perceived control (all multiindicator constructs) were specified as latent variables, and subjective-stress reactivity (residualized change score of repeated single items) and stress exposure (categorical variable) were specified as observed variables. CFI = comparative fit index; TLI = Tucker-Lewis index; RMSEA = root mean square error of approximation; CD = coefficient of determination; SRMR = standardized root mean square residual; ns = not significant. Asterisks indicate significance at *p < .050 and ***p < .001.

In Model 3A, we modeled perceived control as a moderated mediator in the paths from stress exposure (reference = low-stress condition) to stress reactivity and from subjective-stress reactivity to posttest state-anxiety symptoms, respectively. Although Model 3A demonstrated excellent fit by conventional indices (CFI = 1.000, TLI = 1.018, RMSEA = .000, SRMR = .018, p-close = .841), the significant chi-square statistic, χ²(11) = 212.89, p < .001, suggests some model misfit, and the unusually perfect fit indices may indicate overfitting or a near-saturated model. In Model 3A, perceived control significantly moderated the path from stress exposure to subjective-stress reactivity (b = −0.18, SE = 0.05, p = .016) but did not influence the path from subjective-stress reactivity to posttest state-anxiety symptoms (b = −0.08, SE = 0.08, p = .268).

To address the potential issue of overfitting, we estimated an additional model (Model 3B) that omitted the nonsignificant paths from Model 3A. Model 3B reported a lower AIC (5,312.89 vs. 7,261.73) and BIC (5,343.18 vs. 7,295.39) than Model 3A and demonstrated excellent and a more credible model fit, χ²(9) = 212.20, p < .001; CFI = .995, TLI = .989, RMSEA = .035, SRMR = .026, p-close = .528.

In Model 3B, significant direct effects were observed for stress exposure on subjective-stress reactivity (b = 0.59, SE = 0.04, p < .001) such that inclusion in the high-stress condition predicted greater subjective-stress reactivity. As expected, perceived control moderated this path (b = −0.18, SE = 0.07, p = .014), indicating that higher perceived control attenuated increases in subjective stress from baseline to posttask for participants in the experimental condition. This is consistent with our findings from H1. In addition, subjective-stress reactivity significantly predicted posttest state anxiety (b = 0.43, SE = 0.05, p < .001), along with contributions from trait-anxiety symptoms (b = 0.37, SE = 0.05, p < .001). Model 3B also revealed a significant indirect pathway from stress exposure to posttest state anxiety through subjective-stress reactivity (indirect effect: b = 0.24, SE = 0.05, p < .001), supporting the mediating role of subjective-stress reactivity. Likewise, perceived control indirectly reduced state anxiety via lower subjective-stress reactivity (indirect effect: b = −0.06, SE = 0.03, p = .024). A marginal-effects analysis modeling these pathways for low, moderate, and high perceived control revealed that the indirect effect of stress exposure on posttest state-anxiety symptoms was strongest for participants reporting low perceived control (b = 0.30, SE = 0.06, p < .001). For participants reporting moderate perceived control, the indirect effect remained significant but reduced (b = 0.24, SE = 0.05, p < .001). Expectedly, for participants reporting high perceived control, the indirect effect was smallest but still statistically significant (b = 0.18, SE = 0.05, p < .001). These pathways and indirect effects shed light on the roles of stress exposure and perceived control in shaping proximal anxiety responses.

Cross-validation of the SEM

Given the high CFI and TLI statistics, a 10-fold cross-validation was performed to validate the robustness of the findings from Model 3B. Consistent with Model 3B, the cross-validation demonstrated excellent and consistent overall fit across folds (RMSEA: M = .045, SD = .028, minimum = .000, maximum = .099; SRMR: M = .03, SD = .01, minimum = .02, maximum = .05; CD: M = .48, SD = .02, minimum = .45, maximum = .51). Furthermore, the parameter estimates and associated p values demonstrated low variability across folds (see supplementary file “Cross Validation Output” in the study repository on OSF). The close alignment of these results with Model 3B underscores the reliability and generalizability of the model. Low variability across folds further supports the stability of the findings and provides robust evidence for the role of perceived control as a moderated mediator in the relationship between stress exposure and subjective-stress reactivity and its indirect effects on posttest state-anxiety symptoms.

Discussion

In the present study, we examined the role of perceived control in the interplay between stress and state-anxiety symptoms. Building off previous literature, we find that our results offer nuanced clarity regarding the temporal limits and contextual boundaries of the protective nature of perceived control. Several noteworthy findings are discussed below.

Perceived control’s separate relationships with subjective stress and state-anxiety symptoms

Consistent with previous research, we found that perceived control weakened the influence of stress exposure on both changes in subjective stress and state-anxiety symptoms, respectively. This influence was evident only for participants in the high-stress condition; perceived control had no influence on changes in subjective stress or state-anxiety symptoms for participants in the low-stress condition. Thus, although perceived control has been understood as a psychological resource associated with lower overall stress and muted stress responses, our findings suggest that this may be contingent on the degree of subjective stress experienced. It may be the case that perceived control does not play a meaningful role in stress regulation or predict state-anxiety symptoms in low-stress environments. Instead, the protective properties of perceived control may be revealed only in high-stress environments. This comports with Lazarus and Folkman’s (1984) assertion that favorable control beliefs may diminish the perceived intensity of a stressor in large part because of one’s confidence in one’s ability to effectively cope with the stressor. Thus, individuals who believe in their ability to elicit desired outcomes (i.e., participants with high perceived control) may be less reactive to stressors spanning a variety of high-stress domains (Lachman & Weaver, 1998). Extending previous research, however, we find that these results imply that there may be a threshold effect such that environmental stress must surpass a specific level for the benefits of perceived control on subjective stress and state-anxiety symptoms to become evident.

Perceived control as a moderated mediator in the stress–anxiety pathway

Drawing from this, we find that our results also demonstrate the temporal limits of perceived control’s influence on state-anxiety symptoms. Specifically, we found that the influence of perceived control on anxiety was entirely indirect, exerted only through its ability to reduce subjective-stress responses. This provides much-needed clarity on the relationship between perceived control and anxiety symptoms. Although previous research has consistently highlighted the beneficial effects of perceived control on anxiety (Gallagher, Naragon-Gainey, & Brown, 2014; Troy et al., 2013), our findings help explain how this association exists: The effect of perceived control on state anxiety is contingent on its capacity to interrupt the buildup of subjective stress. Beyond this point—once subjective-stress levels have escalated—perceived control appears to play a diminished role in shaping anxiety responses. This threshold-like dynamic is supported by the totality of our findings, which demonstrate a significant moderating effect of perceived control on subjective stress but not on state-anxiety symptoms directly and a fully mediated indirect effect. Beyond this stage, perceived control seems to play a limited role in shaping state-anxiety symptoms. This process can be likened to a balloon being filled with air such that the balloon represents an individual’s capacity to tolerate stress and the air symbolizes the accumulation of subjective stress. Perceived control does not entirely stop the air from entering the balloon; instead, it acts to slow the rate of airflow, thereby delaying the balloon from reaching its critical capacity and ultimately popping. In the same way, perceived control does not prevent subjective stress from increasing state-anxiety symptoms altogether, but it tempers the rate or magnitude of stress buildup. This, in turn, disrupts the stress–anxiety pathway only if deployed early or effectively enough. This insight introduces a limitation for the efficacy of perceived control and underscores the importance of early stress regulation in preventing downstream anxiety.

Clinical and practical implications

These findings have important clinical and practical implications. Considering our finding that perceived control exerts its most significant protective effect in high-stress environments, we suggest that interventions should prioritize addressing perceived control for individuals experiencing chronic or acute stress (e.g., caregivers or individuals undergoing major life transitions, such as job loss or divorce). Such interventions should integrate (a) cognitive-behavioral approaches (e.g., cognitive restructuring) to challenge maladaptive thoughts about controllability and (b) dialectical-behavioral approaches (e.g., mindfulness-based stress reduction) to help individuals remain grounded in the face of overwhelming demands. That said, the process of enhancing perceived control may be more feasible for individuals in low-stress environments (Skinner et al., 1998). In other words, although perceived control seems to be a more valuable asset when experiencing higher stress, it may be feasible to cultivate while experiencing lower stress. Therefore, clinicians may be most effective by prioritizing the development of perceived control as a proactive buffer against future stressors.

This may be particularly beneficial for individuals who present an increased risk for anxiety disorders but do not currently meet diagnostic criteria because fostering a stronger perception of control may fortify individuals against the effect of future stress on downstream anxiety symptoms. In this vein, preventive interventions aimed at fostering perceived control could be incorporated at the population level as well (e.g., community-based programs, professional-development training, and educational settings). These intervention efforts could include self-efficacy training, mentorship, peer support groups, and prompted journaling—all of which have shown efficacy as a complement to treatment (Nasir & Iqbal, 2019; Sohal et al., 2022). They might also include psychoeducation about the temporal relationships between perceived control, subjective stress, and state anxiety. For example, the balloon analogy mentioned above illustrates how perceived control acts as a mechanism that slows the accumulation of subjective stress and disrupts anxiety symptoms: Approaches like this may be an accessible way to facilitate a deeper, practical understanding of therapeutic strategies among individuals who are not so familiar with psychological phenomena. Of course, further research is needed to confirm these notions in the specific context of our findings. Specifically, longitudinal studies may uncover more distal dynamics of the interplay between perceived control, subjective stress, and anxiety symptoms, and experimental studies may further clarify the effects of changes in perceived control on subjective stress and anxiety symptoms. These additional studies will be crucial to enhancing the understanding of the interplay between perceived control, subjective stress, and anxiety symptoms.

Limitations and Future Directions

Although the findings of this study provide valuable insights into the role of perceived control in the stress–anxiety pathway, additional limitations warrant consideration.

First, in the present study, we relied heavily on self-report questionnaires to assess perceived control, subjective stress, and anxiety symptoms. Although widely used and validated, self-report measures are susceptible to response biases (e.g., social-desirability bias or errant self-assessment). Future studies should seek to incorporate physiological (e.g., cortisol levels, heart rate variability) or behavioral (e.g., eye tracking, facial expressions) measures alongside self-reports to offer a more holistic perspective on these findings. Likewise, future studies should investigate whether these findings are applicable across cultures. Because the emphasis of perceived control on well-being may be more prominent in individualistic cultures, findings may differ for data collected in more collectivistic cultures. Given that psychological intervention occurs around the world and aims to improve health and well-being broadly, it is critical that additional work more fully elucidates to what extent and in what groups these results generalize.

Although the experimental manipulation used in the present study distinguished between low-stress and high-stress conditions, it does not fully capture the complexity of real-world stressors. Stressors in everyday life often involve prolonged, multifaceted challenges, but the stressor used in the present study was short-term and task specific. Thus, although we demonstrated the importance of perceived control in mitigating stress escalation during the initial stages of stress appraisal, the temporal dynamics of perceived control’s influence remain underexplored. Key questions for future research include the following: How long do the protective effects of perceived control persist? Can these effects be sustained under prolonged stress? Can perceived control be promoted in the early, low-stress conditions preceding the onset of prolonged, multifaceted, high-stress challenges?

Furthermore, state-anxiety symptoms and part of the stress–reactivity variable were assessed within the same postmanipulation section. Thus, the moderated mediation model should be interpreted as reflecting process-level associations among closely timed responses rather than long-term temporal dynamics. Nevertheless, the experimental manipulation and pre-post residualized change score in subjective stress establish a clear temporal sequence for the mediator, allowing meaningful yet temporally constrained inferences about how the stressor and perceived control jointly shape posttest state-anxiety symptoms. Future studies should confirm the findings in the present study—specifically from the moderated mediation path model—with a study design that more clearly separates stress–reactivity and state-anxiety assessments across multiple postmanipulation time points to more precisely capture temporal ordering.

Along these lines, state-anxiety symptoms were not measured at baseline, precluding causal inference regarding the buffering role of perceived control in the tested relationships between condition assignment and state-anxiety symptoms. In addition, because of IRB constraints, all baseline measures were assessed after participants were made aware that they would be completing an IQ-style cognitive task. This may have influenced baseline scores of the state measures (i.e., subjective stress and confidence in cognitive abilities). Future studies should both include state-anxiety symptoms at baseline and include all baseline measures before the disclosure of study tasks in an IRB-compliant manner. Doing so will elucidate causal links and provide baseline measures that are less influenced by anticipatory stress or task-related expectations.

Finally, these findings suggest a threshold effect in which perceived control is most impactful under higher stress conditions. However, this threshold was not explicitly measured or defined. Future studies should examine the possibility of a threshold effect in the relationship between perceived control and changes in stress, which may include quantifying the threshold of stress necessary for perceived control to be influential and characterizing how this threshold may vary across individuals or stressor types. Likewise, although in the present study we demonstrate that perceived control buffers stress escalation, we do not explore the mechanisms driving this effect. Future studies should explore potential mediators (e.g., emotion regulation, coping strategies employed) to clarify how perceived control exerts its influence on stress appraisals.

Conclusion

The findings of the present study emphasize the importance of perceived control in the stress–anxiety pathway, particularly in high-stress environments. By demonstrating its moderating role in the relationship between stress exposure, subjective stress, and anxiety symptoms, we advance understanding of the contextual boundaries and acute temporal dynamics of perceived control as a protective factor. These findings may also refine existing theoretical frameworks of the cognitive experience of stress and anxiety, underscoring the need for targeted interventions that bolster one’s perceived control. Future research should further explore the long-term implications of perceived control, its role across diverse populations, and the specific mechanisms through which it exerts its influence. These efforts along with the findings of the present study may offer a foundation for more precise and effective clinical and population-level interventions.

Footnotes

Transparency

Action Editor: DeMond M. Grant

Editor: Jennifer L. Tackett

Author Contributions

Christopher J. Davis: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Validation; Visualization; Writing – original draft; Writing – review & editing.

Sydnie R. Spearman: Conceptualization; Methodology; Visualization; Writing – original draft; Writing – review & editing.

Anthony L. Burrow: Conceptualization; Methodology; Supervision; Writing – review & editing.

ORCID iDs

Christopher J. Davis

Sydnie R. Spearman

Anthony L. Burrow

Notes

References

Bandura

(2015). Self-efficacy conception of anxiety. In Schwarzer

Wicklund

(Eds.), Anxiety and self-focused attention (pp. 89–110). Routledge. https://doi.org/10.4324/9781315674766

Barlow

D. H.

(2004). Anxiety and its disorders: The nature and treatment of anxiety and panic. The Guilford Press.

Barré

Brunel

Barthet

Laurencin-Dalicieux

(2017). The visual analogue scale: An easy and reliable way of assessing perceived stress. Quality in Primary Health Care, 1(1), 1–5.

Bhanji

J. P.

Delgado

M. R.

(2014). Perceived control influences neural responses to setbacks and promotes persistence. Neuron, 83(6), 1369–1375.

Biggs

Brough

Drummond

(2017). Lazarus and Folkman’s psychological stress and coping theory. In Cooper

C. L.

Quick

J. C.

(Eds.), The handbook of stress and health: A guide to research and practice (pp. 349–364). John Wiley & Sons. https://doi.org/10.1002/9781118993811

Binfet

J. T.

Passmore

H. A.

Cebry

Struik

McKay

(2018). Reducing university students’ stress through a drop-in canine-therapy program. Journal of Mental Health, 27(3), 197–204.

Boi

Llera

S. J.

(2023). Contrast avoidance prospectively mediates effects of fear of emotional responding, negative problem orientation, and sensitivity to low perceived control on generalized anxiety disorder symptoms. Journal of Anxiety Disorders, 95, Article 102682. https://doi.org/10.1016/j.janxdis.2023.102682

Bollini

A. M.

Walker

E. F.

Hamann

Kestler

(2004). The influence of perceived control and locus of control on the cortisol and subjective responses to stress. Biological Psychology, 67(3), 245–260.

Chorpita

B. F.

Barlow

D. H.

(1998). The development of anxiety: The role of control in the early environment. Psychological Bulletin, 124(1), 3–21. https://doi.org/10.1037/0033-2909.124.1.3

10.

Davis

C. J.

Burrow

A. L.

(2024). Perceived control ameliorates the impact of adverse childhood experiences on downstream mental health. Child Abuse & Neglect, 157, Article 107015. https://doi.org/10.1016/j.chiabu.2024.107015

11.

Davis

C. J.

Levinbook

E. J.

Spearman

S. R.

Burrow

A. L.

(2024). Derailment is associated with increased depressed mood at lower levels of stress: An experimental approach. Journal of Affective Disorders, 363, 320–330.

12.

Daviu

Bruchas

M. R.

Moghaddam

Sandi

Beyeler

(2019). Neurobiological links between stress and anxiety. Neurobiology of Stress, 11, Article 100191. https://doi.org/10.1016/j.ynstr.2019.100191

13.

Delgado

M. R.

Nearing

K. I.

LeDoux

J. E.

Phelps

E. A.

(2008). Neural circuitry underlying the regulation of conditioned fear and its relation to extinction. Neuron, 59(5), 829–838.

14.

Dickerson

S. S.

Kemeny

M. E.

(2004). Acute stressors and cortisol responses: A theoretical integration and synthesis of laboratory research. Psychological Bulletin, 130(3), 355–391. https://doi.org/10.1037/0033-2909.130.3.355

15.

Dijkstra

M. T. M.

Homan

A. C.

(2016). Engaging in rather than disengaging from stress: Effective coping and perceived control. Frontiers in Psychology, 7, Article 1415. https://doi.org/10.3389/fpsyg.2016.01415

16.

Faul

Erdfelder

Buchner

Lang

A. G.

(2009). Statistical power analyses using G* Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41(4), 1149–1160.

17.

Fuentes-García

J. P.

Villafaina

Collado-Mateo

De la Vega

Olivares

P. R.

Clemente-Suárez

V. J.

(2019). Differences between high vs. low performance chess players in heart rate variability during chess problems. Frontiers in Psychology, 10, Article 409. https://doi.org/10.3389/fpsyg.2019.00409

18.

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

19.

Gallagher

M. W.

Naragon-Gainey

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. https://doi.org/10.1007/s10608-013-9587-3

20.

Gao

Assink

Cipriani

Lin

(2017). Associations between rejection sensitivity and mental health outcomes: A meta-analytic review. Clinical Psychology Review, 57, 59–74.

21.

Gilchrist

P. T.

McGovern

G. E.

Bekkouche

Bacon

S. L.

Ditto

(2015). The vasovagal response during confrontation with blood-injury-injection stimuli: The role of perceived control. Journal of Anxiety Disorders, 31, 43–48.

22.

Griffin

Martino

R. J.

LoSchiavo

Comer-Carruthers

Krause

K. D.

Stults

C. B.

Halkitis

P. N.

(2022). Ensuring survey research data integrity in the era of internet bots. Quality & Quantity, 56(4), 2841–2852.

23.

Gruenewald

T. L.

Kemeny

M. E.

Aziz

Fahey

J. L.

(2004). Acute threat to the social self: Shame, social self-esteem, and cortisol activity. Psychosomatic Medicine, 66(6), 915–924.

24.

Gunnar

M. R.

Reid

B. M.

Donzella

Miller

Z. R.

Gardow

Tsakonas

N. C.

Thomas

K. M.

DeJoseph

Bendezú

J. J.

(2021). Validation of an online version of the Trier Social Stress Test in a study of adolescents. Psychoneuroendocrinology, 125, Article 105111. https://doi.org/10.1016/j.psyneuen.2020.105111

25.

Halford

Jonsdottir

I. H.

Eek

(2012). Perceived stress, psychological resources and salivary cortisol. In Kristenson

Garvin

Lundberg

(Eds.), The role of saliva cortisol measurement in health and disease (pp. 67–86). Bentham Science Publishers.

26.

Harrell

F. E.

(2001). Regression modeling strategies: With applications to linear models, logistic regression, and survival analysis (Vol. 608). Springer.

27.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55.

28.

Ilioudi

Lindner

Ali

Wallström

Thunström

A. O.

Ioannou

Anving

Johansson

Hamilton

Falk

Ö.

Steingrimsson

(2023). Physical versus virtual reality–based calm rooms for psychiatric inpatients: Quasi-randomized trial. Journal of Medical Internet Research, 25, Article e42365. https://doi.org/10.2196/42365

29.

Jerusalem

Schwarzer

(2014). Self-efficacy as a resource factor in stress appraisal processes. In Jerusalem

Schwarzer

(Eds.), Self-efficacy (pp. 195–214). Taylor & Francis.

30.

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.

31.

Lazarus

R. S.

Folkman

(1984). Stress, appraisal, and coping. Springer.

32.

Lesage

F. X.

Berjot

Deschamps

(2012). Clinical stress assessment using a visual analogue scale. Occupational Medicine, 62(8), 600–605.

33.

Liang

Liu

Zheng

(2025). Blunted perceived control effect on reward anticipation in trait anxiety. Journal of Affective Disorders, 388, Article 119559. https://doi.org/10.1016/j.jad.2025.119559

34.

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. Journal of Behavioral Medicine, 35, 529–537.

35.

Nasir

Iqbal

(2019). Academic self efficacy as a predictor of academic achievement of students in pre service teacher training programs. Bulletin of Education and Research, 41(1), 33–42.

36.

Postmes

Haslam

S. A.

Jans

(2013). A single-item measure of social identification: Reliability, validity, and utility. British Journal of Social Psychology, 52(4), 597–617.

37.

Gao

Guan

Liu

Yang

(2016). Subjective stress, salivary cortisol, and electrophysiological responses to psychological stress. Frontiers in Psychology, 7, Article 229. https://doi.org/10.3389/fpsyg.2016.00229

38.

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.

39.

Sahib

Chen

Cárdenas

Calear

A. L.

(2023). Intolerance of uncertainty and emotion regulation: A meta-analytic and systematic review. Clinical Psychology Review, 101, Article 102270. https://doi.org/10.1016/j.cpr.2023.102270

40.

Skinner

E. A.

(1996). A guide to constructs of control. Journal of Personality and Social Psychology, 71(3), 549–570.

41.

Skinner

E. A.

Zimmer-Gembeck

M. J.

Connell

J. P.

Eccles

J. S.

Wellborn

J. G.

(1998). Individual differences and the development of perceived control. Monographs of the Society for Research in Child Development, 63(2), 1–231.

42.

Sohal

Singh

Dhillon

B. S.

Gill

H. S.

(2022). Efficacy of journaling in the management of mental illness: A systematic review and meta-analysis. Family Medicine and Community Health, 10(1), Article e001154. https://doi.org/10.1136/fmch-2021-001154

43.

Spielberger

C. D.

Gorsuch

R. L.

Lushene

Vagg

Jacobs

G. J. P. A

. (1983). Manual for the State-Trait Anxiety Inventory. Consulting Psychologists Press.

44.

StataCorp. (2023). Stata statistical software: Release 19. StataCorp LLC.

45.

Tabachnick

B. G.

Fidell

L. S.

Ullman

J. B.

(2019). Using multivariate statistics (Vol. 5). Pearson.

46.

Tolin

D. F.

Lee

Levy

H. C.

Das

Mammo

Katz

B. W.

Diefenbach

G. J.

(2021). Psychophysiological assessment of stress reactivity and recovery in anxiety disorders. Journal of Anxiety Disorders, 82, Article 102426. https://doi.org/10.1016/j.janxdis.2021.102426

47.

Troy

A. S.

Shallcross

A. J.

Mauss

I. B.

(2013). A person-by-situation approach to emotion regulation: Cognitive reappraisal can either help or hurt, depending on the context. Psychological Science, 24(12), 2505–2514.

48.

Turiano

N. A.

Chapman

B. P.

Agrigoroaei

Infurna

F. J.

Lachman

(2014). Perceived control reduces mortality risk at low, not high, education levels. Health Psychology, 33(8), 883–890. https://doi.org/10.1037/hea0000022

49.

Wadsworth

L. P.

Hayes-Skelton

S. A.

(2020). Exploring perceived control, a low-control task, and a brief acceptance intervention in a low and high transdiagnostic anxiety sample. Neurology, Psychiatry and Brain Research, 35, 1–9.

50.

Wadsworth

L. P.

Wessman

Beard

Bjorgvinsson

(2019). Levels of perceived control and treatment response in a brief partial hospital setting. Neurology, Psychiatry and Brain Research, 34, 1–8.

51.

Wen

J. H.

Sin

N. L.

(2022). Perceived control and reactivity to acute stressors: Variations by age, race and facets of control. Stress and Health, 38(3), 419–434.

52.

Zajacova

Lynch

S. M.

Espenshade

T. J.

(2005). Self-efficacy, stress, and academic success in college. Research in Higher Education, 46, 677–706.

53.

Zhao

Lian

Huang

(2021). Exploring emotion regulation and perceived control as antecedents of anxiety and its consequences during Covid-19 full remote learning. Frontiers in Psychology, 12, Article 675910. https://doi.org/10.3389/fpsyg.2021.675910

54.

Zheng

Miao

Gan

(2020). Perceived control buffers the effects of the COVID-19 pandemic on general health and life satisfaction: The mediating role of psychological distance. Applied Psychology: Health and Well-Being, 12(4), 1095–1114.

55.

Zilioli

Imami

Slatcher

R. B.

(2017). Socioeconomic status, perceived control, diurnal cortisol, and physical symptoms: A moderated mediation model. Psychoneuroendocrinology, 75, 36–43.