Perseverative Thinking Questionnaire–Clinical Psychology Trainees (PTQ‑CPT): adaptation and initial validation

Full reference: Dereix-Calonge, I., Ruiz, F. J., Suárez-Falcón, J. C., & Flórez, C. L. (2019). Adapting the Perseverative Thinking Questionnaire for measuring repetitive negative thinking in clinical psychology trainees. Training and Education in Professional Psychology, 13(2), 145-152. https://doi.org/10.1037/tep0000229

Study type: Adaptation and validation / Factor structure study

Background and objectives

Repetitive Negative Thinking (RNT) is a transdiagnostic process identified as a common factor in the onset and maintenance of emotional disorders such as depression, generalized anxiety disorder, social anxiety disorder, posttraumatic stress disorder, and obsessive-compulsive disorder. Although empirical evidence regarding the role of RNT in clinical psychology trainees is limited, these students face multiple stressors during their training, including dealing with patient suffering, lack of support, disagreements with supervisors, diagnostic ambiguity, clinical competence demands, and internal difficulties such as perceived incompetence. Previous research shows that experiential avoidance and values disengagement are associated with depression and burnout in medical students and addiction counselors, but no instrument existed specifically to measure RNT focused on clinical practice. The aim of the present study was to adapt and preliminarily evaluate the psychometric properties of the Perseverative Thinking Questionnaire–Clinical Psychology Trainees (PTQ-CPT), a measure of RNT focused on clinical practice. The authors expected the PTQ-CPT to show (a) good internal consistency, (b) a one-factor structure, (c) medium to strong positive correlations with the original PTQ and measures of experiential avoidance and emotional symptoms, and (d) medium negative correlations with measures of valued living and life satisfaction.

Method

Participants

The sample consisted of 412 undergraduate students from a Colombian university (mean age = 22.74 years, SD = 3.57; 83.5% women). Approximately half of the participants were in their ninth semester (57.3%), while the other half were in their eighth semester (42.7%). Ninth-semester students were at the beginning of their clinical practice, while eighth-semester students were receiving training in clinical psychology to begin practice in the next semester. Unlike other countries such as the United States, Colombian law permits undergraduate psychology students to receive clinical psychology training and attend to clients under professional supervision. Nearly all participants were single (94.4%). Forty percent had received some form of psychological or psychiatric treatment in the past, but only 3.7% were in treatment during the study (1.2% were taking psychotropic medication).

Instrument(s) under study

Perseverative Thinking Questionnaire–Clinical Psychology Trainees (PTQ-CPT). The authors modified the 15 items of the original PTQ (Ehring et al., 2011) to resemble RNT occurring in clinical psychology practice. Instructions were reformulated to normalize the difficulties participants might experience. Most changes consisted of specifying RNT engagement regarding clinical practice issues (e.g., "My thoughts are repetitive" became "My thoughts about the clinical practice are repetitive"). Four experts (three with experience as clinical supervisors and notions of psychometrics, and one with exclusive experience in psychometrics; all Spanish speakers) qualitatively evaluated the items using a 4-point Likert scale (4 = very; 1 = nothing) according to representativeness, understandability, interpretability, and clarity. Aiken's V and 95% confidence intervals were used to determine item adequacy, with a suggested cutpoint of less than .50 for inadequacy. Six items were eliminated and some were modified according to expert suggestions, resulting in a final 9-item version. The final version uses a Likert-type response scale, though the exact scoring range is not specified in the PDF.

Other outcome measures

Perseverative Thinking Questionnaire (PTQ). A 15-item 5-point Likert scale (4 = almost always; 0 = never) instrument designed to evaluate the tendency to engage in RNT when facing negative experiences or problems. The PTQ is a content-independent measure. It showed a hierarchical factor structure with a higher-order factor representing general RNT and three lower-order factors: (a) core features of RNT, (b) unproductiveness of RNT, and (c) mental capacity captured by RNT. In this study, α = .95.

Acceptance and Action Questionnaire–II (AAQ-II). A general measure of experiential avoidance consisting of 7 items on a 7-point Likert scale (7 = always true; 1 = never true). The Spanish version showed good psychometric properties and one-factor structure in Colombia. In this study, α = .91.

Depression, Anxiety, and Stress Scales–21 (DASS-21). A 21-item 4-point Likert-type scale (3 = applied to me very much or most of the time; 0 = did not apply to me at all) with three subscales: Depression, Anxiety, and Stress. It showed good psychometric properties and hierarchical factor structure in Colombia. In this study, α = .91 (total scale), .87 (Depression), .77 (Anxiety), and .83 (Stress).

General Health Questionnaire–12 (GHQ-12). A 12-item 4-point Likert-type scale frequently used as a screening tool for psychological disorders. It showed excellent psychometric properties in Colombia and one-factor structure. In this study, α = .86.

Valuing Questionnaire (VQ). A 10-item 6-point Likert scale (6 = completely true; 0 = not at all true) designed to assess general valued living during the past week. It has two subscales: Progress (enactment of values, with clear awareness of what is personally important and perseverance) and Obstruction (disruption of valued living due to avoidance of unwanted experience and distraction from values). In this study, α = .78 (Progress) and .76 (Obstruction).

Satisfaction With Life Scale (SWLS). A 5-item 7-point Likert-type scale (7 = strongly agree; 1 = strongly disagree) measuring self-perceived well-being. It showed good psychometric properties in Colombia. In this study, α = .85.

Data analysis

Following qualitative expert evaluation, a cross-validation study was conducted to analyze the factor structure of the PTQ-CPT. This procedure involved splitting the recruited sample into two random subsamples of approximately equal size and conducting an Exploratory Factor Analysis (EFA) with the first subsample and a Confirmatory Factor Analysis (CFA) with the second subsample. The sample was split using SPSS 19.

The EFA in the first subsample was conducted with Factor 10.5 software using unweighted least squares extraction method with Direct Oblimin rotation, using polychoric correlations. The number of dimensions was determined through the optimal implementation of Parallel Analysis (PA) based on minimum rank factor analysis. Unidimensionality assessment was conducted using Unidimensional Congruence (UniCo), Explained Common Variance (ECV), and Mean of Item Residual Absolute Loadings (MIREAL) indexes. Cronbach's coefficient alpha with 95% confidence intervals was calculated, and corrected item-total correlations were obtained.

For the CFA in the second subsample, a robust Weighted Least Squares (Robust WLS) estimation method using polychoric correlations was adopted, using LISREL (Version 8.71). The chi-square test and the following goodness-of-fit indexes were computed: (a) Root Mean Square Error of Approximation (RMSEA), (b) Comparative Fit Index (CFI), and (c) Non-Normed Fit Index (NNFI). According to interpretation criteria, RMSEA values of .10 represent good fit and values below .05 represent very good fit; CFI and NNFI values above .90 indicate well-fitting models, and those above .95 represent very good fit.

Finally, Pearson correlations between the PTQ-CPT and other scales were calculated to assess convergent construct validity.

Results

Following qualitative item evaluation by experts, six items of the PTQ-CPT were eliminated. Specifically, items 2, 10, 12, and 13 were eliminated due to obtaining low Aiken's V scores. Item 4 was eliminated because it assumed clinical practice had already begun. Item 7 was eliminated due to excessive similarity with item 3. Four items (1, 8, 11, and 14) were slightly modified according to expert suggestions, improving their Aiken's V scores. The final PTQ-CPT version consisted of 9 items.

All items showed good discrimination, with corrected item-total correlations ranging from .66 to .78. Cronbach's alpha coefficient was .93 (95% CI [.91, .94]), indicating excellent internal consistency. In the exploratory factor analysis with the first subsample (N = 204; mean age = 22.77 years, SD = 3.75; 42.8% in eighth semester; 85.4% women), Bartlett's test was statistically significant (χ² (36) = 1230.2, p < .001), and the Kaiser-Meyer-Olkin test was very good (.92). Parallel Analysis suggested extracting a single factor, which explained 70.4% of the variance (eigenvalue = 6.34). Factor loadings were high for all items, ranging from .73 (item 1) to .88 (item 5). The indexes of UniCo (.99), ECV (.92), and MIREAL (.21) strongly supported the unidimensionality of the PTQ-CPT.

In the confirmatory factor analysis conducted with the second subsample (N = 203; mean age = 22.70 years, SD = 3.39; 43.2% in eighth semester; 81.6% women), the overall fit of the one-factor model was good: χ² (27) = 49.409, p < .01; RMSEA = .064, 90% CI [.035, .092]; CFI = .99; NNFI = .99. The fit indexes exceeded acceptance criteria, demonstrating very good fit of the one-factor model.

Regarding validity based on relationships with other variables, the PTQ-CPT showed theoretically coherent correlations with other relevant constructs. The instrument exhibited medium to strong positive correlations with the PTQ (r = .50, p < .001), indicating a robust relationship with general RNT. It also showed medium to strong positive correlations with experiential avoidance as measured by the AAQ-II (r = .49, p < .001), supporting the conceptualization of RNT as an experiential avoidance strategy. The PTQ-CPT showed significant positive correlations with emotional symptoms as measured by DASS-21 total (r = .52, p < .001), Depression (r = .42, p < .001), Anxiety (r = .43, p < .001), Stress (r = .48, p < .001), and GHQ-12 (r = .50, p < .001), confirming its relationship with emotional psychopathology. The instrument presented medium-magnitude negative correlations with measures of valued living and satisfaction, showing negative correlations with Values Progress (VQ; r = -.28, p < .001), Values Obstruction (VQ; r = .50, p < .001), and Satisfaction With Life Scale (r = -.27, p < .001), suggesting that RNT focused on clinical practice is associated with difficulties in values enactment and lower life satisfaction.

Discussion and conclusions

This study presented the adaptation of the PTQ for clinical psychology trainees (PTQ-CPT) and evaluation of its preliminary psychometric properties. The authors modified the 15 original items to reflect RNT in the context of clinical training, which were evaluated by experts, and six items were eliminated, resulting in a 9-item version. This version was administered to a large sample of undergraduate students who were conducting or about to begin their clinical practice.

Results demonstrated solid psychometric properties of the PTQ-CPT. Internal consistency was excellent (α = .93), supporting instrument reliability. The cross-validation study provided strong support for the unidimensionality of the measure, with a one-factor structure clearly demonstrated in both the EFA and CFA. The fit indexes of the confirmatory model were excellent (CFI = .99, NNFI = .99, RMSEA = .064), indicating that the model fits the observed data very well. Factor loadings were consistently high (range .73 to .88), reflecting substantial contributions of each item to the general factor.

Regarding convergent validity, the PTQ-CPT showed a pattern of correlations coherent with the underlying theoretical framework. Significant correlations with the original PTQ (r = .50) validate that the instrument captures aspects of general RNT, although with specificity for the clinical context. Correlations with experiential avoidance (AAQ-II, r = .49) reinforce the theoretical proposal that RNT functions as an avoidance strategy. Medium to strong correlations with emotional symptoms (DASS-21 total r = .52, GHQ-12 r = .50) establish validity related to emotional criteria. Negative correlations with life satisfaction (r = -.27) and values progress (r = -.28) support the construct that RNT interferes with well-being and values enactment, key concepts in the Contextual Behavioral Science perspective. The authors concluded that the PTQ-CPT is a valid and reliable measure of RNT focused on clinical practice and could be used to explore the role of RNT in the stress and learning difficulties experienced by clinical psychology trainees.

Significance and Contribution

This study makes a significant contribution to the measurement of repetitive negative thinking in clinical psychology training contexts. By providing an adaptation of the PTQ specifically developed for clinical psychology students, the study offers a contextualized assessment tool that recognizes repetitive negative thinking as a transdiagnostic process relevant to this population. The validation using rigorous cross-validation methodology (exploratory and confirmatory factor analysis in independent subsamples) demonstrates the soundness of the instrument. The study contributes to the field of psychological measurement in professional training contexts, facilitating the assessment of repetitive negative thinking related to clinical practice and supporting both research and clinical application in Spanish-speaking populations of clinical psychology students.