Psychological Inflexibility Mediates the Effects of Self-Efficacy and Anxiety Sensitivity on Worry

Full reference: Ruiz, F. J. (2014). Psychological inflexibility mediates the effects of self-efficacy and anxiety sensitivity on worry. Spanish Journal of Psychology, 17, e3. https://doi.org/10.1017/sjp.2014.3

Study type: Cross-sectional correlational study with mediation analysis

Background and objectives

Pathological worry constitutes a central feature of generalized anxiety disorder (GAD), a chronic disorder associated with significant functional impairment, high rates of comorbidity, and difficulty in treatment. Previous research has identified several factors associated with the development and maintenance of pathological worry, including general self-efficacy (GSE; beliefs about the ability to handle a wide range of stressors) and anxiety sensitivity (AS; the tendency to perceive anxious states as aversive and harmful).

The present study is grounded in the Acceptance and Commitment Therapy (ACT) model, which emphasizes the role of psychological inflexibility (PI; the dominance of private experiences over chosen values and contingencies in guiding action) as a central mechanism in the development and maintenance of psychopathology, including generalized anxiety. The ACT model of GAD proposes that the inability to be in contact with fear and anxiety leads to the use of worry as an experiential avoidance strategy that is negatively reinforced in the short term.

The study objectives were: (1) to examine zero-order correlations among general self-efficacy, anxiety sensitivity, psychological inflexibility, and pathological worry; (2) to determine whether GSE and AS remained independent predictors of worry even after controlling for each other; and (3) to investigate whether psychological inflexibility completely mediated the effects of both variables on pathological worry.

Method

Participants

The sample consisted of 132 nonclinical participants with an age range of 18 to 69 years (M = 33.56, SD = 12.88) recruited from the province of Almería (Southeast of Spain). Sixty-three percent were women. Regarding educational level, 65% had completed or were currently pursuing university coursework, 20% had completed mid-level studies, and 15% had completed primary studies.

Design

A cross-sectional correlational design was employed. Participants completed a packet of self-report questionnaires during a single assessment session. Participants were recruited from undergraduate psychology and pedagogy students and their personal contacts. All participants signed a general informed consent document and were subsequently debriefed regarding the study aims.

Intervention / Conditions

No experimental intervention was administered. The study was purely observational. Self-report measures were collected at a single time point without experimental manipulation.

Outcome measures

Penn State Worry Questionnaire (PSWQ; Meyer, Miller, Metzger, & Borkovec, 1990): A 16-item, 5-point Likert scale self-report instrument designed to evaluate the permanent and nonspecific degree of worry that characterizes GAD. Example items include "My worries overwhelm me" and "I know I should not worry about things, but I just cannot help it." Internal consistency is high (alpha between .93 and .95). The Spanish translation by Sandín, Chorot, Valiente, and Lostao (2009) was administered, which showed mean scores of 49.8 (SD = 12.6).

General Self-Efficacy Scale (GSES; Schwarzer & Jerusalem, 1995): A 10-item, 4-point Likert scale that measures beliefs about the ability to cope with a wide range of stressors. Example items are "I can always manage to solve difficult problems if I try hard enough" and "I am confident that I could deal efficiently with unexpected events." The GSES shows good internal consistency (alpha values between .79 and .93) and a one-factor solution. The Spanish translation by Baessler and Schwarzer (1996) was used. The mean score found in several studies is approximately 31.

Anxiety Sensitivity Index (ASI; Peterson & Reiss, 1992): A 16-item, 5-point Likert scale that measures anxiety sensitivity. Participants rate the extent to which they expect negative consequences from a variety of anxiety-related experiences, including physical concerns (e.g., "When I notice my heart beating rapidly, I worry that I might have a heart attack"), cognitive concerns (e.g., "When I cannot keep my mind on a task, I worry that I might be going crazy"), and social concerns (e.g., "It is important to me not to appear nervous"). The Spanish adaptation of the ASI has good psychometric properties in clinical and normal populations. Mean scores in nonclinical populations were 19.7 (SD = 9.2) for the complete scale, 8.8 (SD = 5.3) for physical concerns, 3.5 (SD = 3.3) for cognitive concerns, and 7.2 (SD = 2.9) for social concerns.

Acceptance and Action Questionnaire - II (AAQ-II; Bond et al., 2011): A general measure of experiential avoidance and psychological inflexibility consisting of 7 items rated on a 7-point Likert scale. Items reflect unwillingness to experience unwanted emotions and thoughts (e.g., "I am afraid of my feelings," "I worry about not being able to control my worries and feelings") and the inability to be in the present moment and carry out value-directed actions when experiencing psychological events that could undermine them (e.g., "My painful experiences and memories make it difficult for me to live a life that I would value," "My painful memories prevent me from having a fulfilling life," "Worries get in the way of my success"). The Spanish translation by Ruiz, Langer, Luciano, Cangas, and Beltrán (2013) was used, which showed a one-factor solution, good internal consistency (mean α = .88), and external validity. The mean score found in nonclinical samples was 21.22 (SD = 7.76).

Primary dependent variable: Penn State Worry Questionnaire (pathological worry).

Data analysis

Statistical analyses were conducted in stages. First, zero-order correlations among all variables were computed (PSWQ, GSES, ASI, AAQ-II). Second, partial correlations were calculated to explore whether GSE and AS remained associated with pathological worry after controlling for each other. Third, two mediation analyses were conducted using the nonparametric bootstrapping procedure for estimating direct and indirect effects described by Preacher and Hayes (2004). In each mediation analysis, PSWQ scores served as the dependent variable, AAQ-II scores as the proposed mediating variable, and age, gender, and educational level were entered as covariates. GSES and ASI scores served as independent variables, respectively, in the first and second mediation analyses. To control for each other, ASI scores were entered as an additional covariate in the mediation model for GSE, and conversely GSES scores were used as a covariate in the mediation analysis for AS. Mediation was deemed significant if the 95% bias-corrected and accelerated (BCa) bootstrap confidence intervals for the indirect effects based on 20,000 bootstrapped samples did not include zero. Additionally, effect sizes of psychological inflexibility as a mediator were calculated using the completely standardized indirect effect (ab_cs; Preacher & Kelley, 2011).

Results

Preliminary analyses and zero-order correlations

Participants obtained mean scores on pathological worry (PSWQ) of 50.5 (SD = 11.7), general self-efficacy (GSES) of 30.2 (SD = 5.6), total anxiety sensitivity (ASI) of 23.6 (SD = 13.2), and psychological inflexibility (AAQ-II) of 21.0 (SD = 7.9). Very good internal consistencies were found for all scales, with Cronbach's alphas of .92 for PSWQ, .87 for GSES, .92 for total ASI, and .86 for AAQ-II, except for social concerns of the ASI (α = .62).

Zero-order correlations showed that psychological inflexibility was the variable with the highest correlation with pathological worry (r = .56), followed by anxiety sensitivity (r = .40) and general self-efficacy (r = -.29). The ASI subscales also showed significant correlations with PSWQ (r = .34, .35, and .38 respectively for physical, cognitive, and social concerns).

Partial correlations

Partial correlations indicated that GSE remained significantly correlated with pathological worry after controlling for AS (r = -.19, p = .027). Conversely, AS also remained significantly correlated with worry after controlling for GSE (r = .35, p < .001). The ASI subscales remained significantly correlated with PSWQ scores (.29, .29, and .34 respectively for physical, cognitive, and social concerns).

Mediation analysis of the effect of GSE on pathological worry through psychological inflexibility

Results showed that general self-efficacy significantly predicted psychological inflexibility (path a: B = -.539, SE = .123, p < .001). GSE also significantly predicted pathological worry even with AS controlled as an additional covariate along with age, gender, and educational level (path c or total effect: B = -.474, SE = .19, p < .05). However, this effect became nonsignificant when psychological inflexibility was included in the model (path c' or direct effect: B = -.131, SE = .188, p = .486). The indirect effect of GSE on pathological worry through PI was significant (path ab), with a point estimate of -.343 (SE = .101) and a 95% BCa confidence interval of -.580 to -.177. The effect size of this indirect effect was medium (ab_cs = -.162, 95% BCa CI of -.278 to -.084). In conclusion, the mediation analysis revealed that psychological inflexibility completely mediated the relationship between GSE and pathological worry.

Mediation analyses of the effect of AS on pathological worry through psychological inflexibility

Regarding total AS, the first regression analysis showed that it significantly predicted psychological inflexibility (path a: B = .174, SE = .051, p < .001). Likewise, total AS significantly predicted pathological worry even when controlling for GSE (path c or total effect: B = .238, SE = .079, p < .01). However, this effect became nonsignificant when controlling for psychological inflexibility (path c' or direct effect: B = .127, SE = .076, p = .095). The indirect effect of total AS on pathological worry through psychological inflexibility was significant (path ab: point estimate = .111, SE = .045, 95% BCa CI of .034 to .210). The effect size of the indirect effect was medium (ab_cs = .125, 95% BCa CI of .039 to .237).

Similar results were obtained with physical and cognitive concerns of AS. Both factors significantly predicted psychological inflexibility (paths a; physical concerns: B = .265, SE = .091, p < .01; cognitive concerns: B = .530, SE = .149, p < .001) as well as pathological worry (paths c or total effects; physical concerns: B = .333, SE = .141, p < .05; cognitive concerns: B = .511, SE = .235, p < .05). However, these latter effects became nonsignificant after controlling for psychological inflexibility (paths c' or direct effects; physical concerns: B = .157, SE = .132, p = .238; cognitive concerns: B = .153, SE = .225, p = .499), whereas the indirect effects of physical and cognitive concerns on worry through psychological inflexibility were significant (paths ab; physical concerns: B = .176, SE = .070, 95% BCa CI of .059 to .337; cognitive concerns: point estimate = .358, SE = .165, 95% BCa CI of .045 to .671). Effect sizes were medium (physical concerns: ab_cs = .113, 95% BCa CI of .037 to .214; cognitive concerns: ab_cs = .137, 95% BCa CI of .017 to .256).

Finally, social concerns of AS significantly predicted both psychological inflexibility (path a: B = .426, SE = .211, p < .05) and pathological worry (path c or total effect: B = 1.147, SE = .311, p < .001). The final regression analysis showed that, although social concerns still significantly predicted pathological worry after controlling for psychological inflexibility (path c': B = .877, SE = .287, p = .003), the indirect effect of social concerns through psychological inflexibility on worry was also significant (path ab: point estimate = .270, SE = .147, 95% BCa CI of .023 to .601) with a small-medium effect size (ab_cs = .074, 95% BCa CI of .006 to .165).

In conclusion, the mediation analyses revealed that the effects of AS and all its dimensions on pathological worry were significantly mediated by psychological inflexibility.

Discussion and conclusions

This study examined the hypothesis that the effects of GSE and AS on pathological worry would be mediated by psychological inflexibility. As predicted, even when controlling for each other, both GSE and AS were related to pathological worry. Mediation analyses showed that psychological inflexibility completely mediated the effects of GSE and AS on pathological worry. With regard to specific factors of AS, psychological inflexibility also fully mediated the effects of both physical and cognitive concerns on worry. A different pattern of results was obtained for social concerns of AS, where psychological inflexibility only partially mediated the effect on pathological worry.

The study presents important limitations. First, the cross-sectional design does not allow determination of whether changes in predictor variables preceded changes in the criterion variable, preventing the assumption of causal relationships. Relations among variables could be bidirectional. Second, because all data were obtained using self-report measures, relationships among variables might be artificially inflated. Third, as the sample consisted of nonclinical participants, the generalizability of the findings may be limited. Fourth, social concerns of the ASI showed poor internal consistency (α = .62), suggesting that future research should use the Anxiety Sensitivity Index-Revised (ASI-R), which appears to resolve this problem.

Despite these limitations, the results provide important empirical support for the ACT model of GAD. The model emphasizes the central role of experiential avoidance and psychological inflexibility in the development and maintenance of GAD. From this perspective, individuals with low levels of GSE may have a higher tendency to perceive uncertain future outcomes in response to demanding events, while individuals with high AS may be more concerned about probable stressful situations. Both types of individuals are likely to develop pathological levels of worry to the extent that they are psychologically inflexible, rigidly using worry as an experiential avoidance strategy in response to fear and anxiety.

Significance and contribution

This study contributes to the field of anxiety psychopathology by elucidating how low general self-efficacy and high anxiety sensitivity affect pathological worry. The findings demonstrate that psychological inflexibility operates as the central mediator through which other psychological risk factors exert their effects on worry. This work integrates concepts from distinct theoretical traditions—the cognitive model of depression and contextual behavioral science—and provides a transdiagnostic understanding of how mechanisms of acceptance and psychological flexibility operate in the etiology of generalized anxiety.

---