Development and initial validation of the Generalized Pliance Questionnaire

Full reference: Ruiz, F. J., Suárez-Falcón, J. C., Barbero-Rubio, A., & Flórez, C. L. (2019). Development and initial validation of the Generalized Pliance Questionnaire. Journal of Contextual Behavioral Science, 12, 189–198. https://doi.org/10.1016/j.jcbs.2018.03.003

Study type: Scale development and initial validation

Background and objectives

Empirical research on functional classes of rule-governed behavior has been scarce, likely due to the absence of valid behavioral or self-report measures. Generalized pliance is defined as a functional class of rule-governed behavior in which the source of reinforcement is social, characterized by actions directed toward obtaining social approval as the primary life reinforcer, resulting in reduced sensitivity to other sources of stimulus control. According to Relational Frame Theory (RFT), pliance is the first functional class of rule-governed behavior to develop, and its excessive generalization can lead to insensitivity to direct contingencies and rigid patterns of experiential avoidance, being identified as a risk factor for psychopathology development.

Although the terms pliance and tracking were coined in 1982, there is limited experimental evidence that these concepts refer to distinct functional classes of behavior. An important limitation is the lack of valid self-report measures explicitly designed to measure generalized pliance. This study presents the development and initial validation of the Generalized Pliance Questionnaire (GPQ) to fill this gap in the literature and facilitate research on functional classes of rule-governed behavior.

Method

Participants

The study included three independent samples with a total of 2,127 participants. Study 1: 130 undergraduates (age range: 18–46 years, M = 22.58, SD = 5.09) from a university in southern Spain. 77% were Psychology majors; the remaining 23% studied Teaching, Law, or English Philology. 59% were women. Only 3.1% reported being in psychological or psychiatric treatment, and 3.8% were taking psychotropic medication.

Study 2: 410 undergraduates (83.7% women) from a Colombian university with age range 19–49 years (M = 22.74, SD = 3.58). All were Psychology majors. 3.7% reported being in psychological or psychiatric treatment, and 1.2% were taking psychotropic medication.

Study 3 included three Colombian samples: (a) 762 undergraduates (62% women) with age range 18–63 years (M = 21.16, SD = 3.76) from seven universities in Bogotá. 46% were Psychology undergraduate students; other majors included Law, Engineering, Mathematics, and Physics. 4.3% reported being in psychological or psychiatric treatment, and 2.9% were taking psychotropic medication; (b) 724 participants (74.4% females) with age range 18–88 years (M = 26.11, SD = 8.93). 17.8% had primary/mid-level education, 63.8% were undergraduates or college graduates, and 18.4% had or were pursuing graduate degrees. All were Colombian and completed an anonymous online survey. 8.4% reported being in psychological or psychiatric treatment, and 5.4% were using psychotropic medication; (c) 101 patients (52 women) with age range 18–67 years (M = 32.22, SD = 12.09), suffering from emotional disorders (67.3% depression and anxiety disorders) or sexual disorders (32.7%). 5% reported taking psychotropic medication.

Instruments under study

Generalized Pliance Questionnaire (GPQ). In Study 1, a 38-item version (GPQ-38) was developed based on an initial pool of 77 items written by five RFT expert researchers. Two additional RFT experts rated quality, retaining 38 high-quality items. Study 1 resulted in reduction to 18 items (GPQ-18) based on corrected item-total correlations and criterial validity. Study 2 produced a shorter 9-item version (GPQ-9). A 7-point Likert scale was used (7 = always true; 1 = never true).

Other outcome measures

Acceptance and Action Questionnaire-II (AAQ-II; Bond et al., 2011; Spanish translation by Ruiz et al., 2013): 7 items, 7-point Likert scale, measures experiential avoidance. Positive correlations with GPQ were expected.

Beck Depression Inventory-II (BDI-II; Beck, Steer, & Brown, 1996; Spanish translation by Sanz, Perdigón, & Vázquez, 2003): 21 items, 0–3 scale, measures depressive symptoms in adolescents and adults. Positive correlations with GPQ were expected.

Dysfunctional Attitude Scale-Revised (DAS-R; de Graaf, Roelofs, & Huibers, 2009; Spanish translation by Ruiz et al., 2015): 17 items, 7-point Likert scale, two factors (perfectionism/performance evaluation and dependency). Positive correlations with GPQ were expected.

Cognitive Fusion Questionnaire (CFQ; Gillanders et al., 2014; Spanish translation by Ruiz et al., 2017): 7 items, 7-point Likert scale, measures cognitive fusion. Medium to strong positive correlations with GPQ were expected.

Valuing Questionnaire (VQ; Smout, Davies, Burns, & Christie, 2014; Spanish translation by Ruiz et al., submitted): 10 items, 6-point Likert scale, two subscales (Progress and Obstruction). Negative to medium correlations with Progress and medium to strong positive correlations with Obstruction were expected.

Mindfulness Attention Awareness Scale (MAAS; Brown & Ryan, 2003; Spanish translation by Soler et al., 2012): 15 items, 6-point Likert scale, measures attention during tasks. Small to medium negative correlations with GPQ were expected.

Depression, Anxiety, and Stress Scales-21 (DASS-21; Lovibond & Lovibond, 1995; Spanish translation by Daza, Novy, Stanley, & Averill, 2002): 21 items, 4-point Likert scale, three subscales (Depression, Anxiety, Stress). Medium to strong positive correlations with GPQ were expected.

General Health Questionnaire-12 (GHQ-12; Goldberg & Williams, 1988; Spanish translation by Rocha et al., 2011): 12 items, 4-point Likert scale, screening for psychological disorders. Medium to strong positive correlations with GPQ were expected.

Satisfaction with Life Scale (SWLS; Diener, Emmons, Larsen, & Griffin, 1985; Spanish translation by Atienza, Pons, Balaguer, & García-Merita, 2000): 5 items, 7-point Likert scale, measures self-perceived well-being. Medium to large negative correlations with GPQ were expected.

Data analysis

Study 1: Corrected item-total correlations were analyzed to identify low-functioning items. GPQ-18 was retained based on criteria: (a) expert selection, (b) maximization of reliability, and (c) maximum criterial validity according to significant correlations with AAQ-II, BDI-II, and DAS-R.

Study 2: Exploratory factor analysis (EFA) using Factor 10.5 software with unweighted least squares (ULS) extraction, Direct Oblimin rotation, polychoric correlations, and Hot-Deck multiple imputation for missing values. Number of dimensions determined by optimal parallel analysis based on minimum rank factor analysis. Unidimensionality assessed using Unidimensional Congruence (UniCo), Explained Common Variance (ECV), and Mean of Item Residual Absolute Loadings (MIREAL).

Study 3: Confirmatory factor analysis (CFA) using LISREL© (version 8.71) with robust weighted least squares (Robust WLS), polychoric correlations. Fit indices evaluated: chi-square, root mean square error of approximation (RMSEA), comparative fit index (CFI), non-normed fit index (NNFI), and p-value for test of close fit (PCLOSE). RMSEA < .05 criteria for very good fit and < .10 for good fit; CFI and NNFI > .95 for very good fit and > .90 for acceptable fit. Pearson correlations to assess convergent validity with other scales.

Results

Study 1 resulted in a preliminary 18-item version (GPQ-18) with Cronbach's alpha of .90. Study 2's EFA revealed the GPQ-18 is unidimensional (UniCo = .98, ECV = .90, MIREAL = .22), accounting for 56.3% of variance (eigenvalue = 10.13). Factor loadings were high, ranging from .56 (Item 17) to .87 (Item 13). A shorter 9-item version (GPQ-9) was developed, accounting for 72.6% of variance (eigenvalue = 5.94), with factor loadings ranging from .65 (Item 2) to .90 (Item 6), confirming unidimensionality (UniCo = .99, ECV = .91, MIREAL = .20).

Study 3's CFA with combined samples 3–5 (N = 1,587) showed good fit for the one-factor model in GPQ-18: χ²(135) = 784.35, p < .01; RMSEA = .055 [90% CI .051, .058], CFI = .96, NNFI = .96. PCLOSE was not statistically significant at p < .01 (p = .012), indicating close fit. Fit for GPQ-9 was slightly worse but acceptable: χ²(27) = 327.32, p < .01; RMSEA = .084 [90% CI .076, .092], CFI = .97, NNFI = .95, although PCLOSE was significant (p < .001).

Alpha coefficients were excellent: GPQ-18 of .93, .95, and .97 for samples 3–5 respectively; GPQ-9 of .91, .94, and .95. Correlations between GPQ-18 and GPQ-9 were nearly perfect (.95, .98, .98). Regarding convergent validity, the GPQ showed strong positive correlations with experiential avoidance (AAQ-II) and cognitive fusion (CFQ), medium to strong positive correlations with emotional symptoms (DASS-21 and GHQ-12), and small to medium negative correlations with life satisfaction (SWLS) and mindfulness (MAAS). Correlations with obstruction in valued living (VQ-Obstruction) were medium to strong and positive, while with progress (VQ-Progress) were negative.

Discussion and conclusions

The study successfully describes the development and initial validation of the GPQ as a valid and reliable measure of generalized pliance. Two psychometrically robust versions were obtained: the 18-item GPQ-18 (alphas between .93–.97) and the 9-item GPQ-9 (alphas between .91–.95). Both are unidimensional, as confirmed by EFA, CFA, and unidimensionality analyses. The pattern of correlations with other measures is theoretically coherent with the conceptualization of generalized pliance within RFT: pliance generalization was expected and found to relate strongly with experiential avoidance, cognitive fusion, emotional symptoms, and low life satisfaction.

The authors acknowledge limitations: disproportionate female representation, small clinical sample, exclusive use of self-reports (which may have inflated correlations), and psychometric properties reported only in Colombia. A subsequent study (O'Connor, Byrne, Ruiz, & McHugh, submitted) used an English version of the GPQ and found strong correlations with two behavioral tasks measuring insensitivity to contingencies (r = .56 and .44 for WCST and Iowa Gambling Task, respectively), supporting the GPQ's criterion validity. Future research is recommended with gender-balanced samples, larger clinical samples, cross-cultural validation, correlations with behavioral insensitivity tasks, and development of context-specific pliance measures for specific situations (couple relationships, work, etc.).

Significance and contribution

This study contributes significantly to the field of rule-governed behavior measurement by developing and validating the Generalized Pliance Questionnaire as a valid, reliable, and theoretically grounded instrument. By providing both an 18-item version and an abbreviated 9-item version with excellent psychometric properties (alphas ranging from .91 to .97), the work facilitates research on functional classes of rule-governed behavior. The coherent patterns of correlations with psychological inflexibility, emotional symptoms, and life satisfaction confirm the theoretical relevance of generalized pliance, enabling future longitudinal research on its role in psychopathology and treatment response.