Psychometric properties of the Valuing Questionnaire in Colombian samples

Background and objectives

The Valuing Questionnaire (VQ) is one of the most widely used and psychometrically robust instruments for measuring valued living within the context of Acceptance and Commitment Therapy (ACT). This study examined the psychometric properties of the Spanish version of the Valuing Questionnaire in Colombian clinical and nonclinical samples, with the objective of establishing the validity and reliability of the instrument for use in research and clinical practice in Spanish-speaking populations.

Method

Participants

The study included three independent samples totaling 1,820 participants:

Sample 1 (Undergraduates): 762 participants (62% female), with age range 18–63 years (M = 21.16, SD = 3.76). Participants were recruited from seven universities in Bogotá, primarily from psychology faculties, though students from engineering, medicine, business, law, theology, communication, and philosophy were also included. Of the total, 4.3% were currently receiving psychological or psychiatric treatment, and 26% had received treatment in the past. Only 2.9% were on psychotropic medication.

Sample 2 (General Population): 724 participants (74.4% female), with age range 18–88 years (M = 26.11, SD = 8.93). Educational level was distributed as follows: 17.8% completed compulsory education, 37.3% completed secondary school, 28.9% were undergraduate students or college graduates, and 18.4% completed or were enrolled in graduate programs. Participants were recruited through online surveys distributed via social media. Of the total, 5.4% were on psychotropic medication.

Sample 3 (Clinical): 334 participants (66.8% female), with age range 18–67 years (M = 28.59, SD = 11.09). All participants were evaluated at a clinical psychology center of a Colombian university that provides inexpensive psychological intervention to the general population. All reported clinical levels of mental health difficulties, with 6% receiving psychotropic medication at the time of the study.

Instrument evaluated: Valuing Questionnaire (VQ)

The VQ is a self-report questionnaire consisting of 10 items rated on a 7-point Likert-type scale (6 = completely true; 0 = not at all true). The instrument measures the extent to which individuals enact their personal values during the past week, without reference to specific life domains. The VQ has demonstrated a two-factor structure in multiple studies: (a) Progress, reflecting commitment and advancement in valued life areas, and (b) Obstruction, reflecting disruption of valued living due to avoidance of unwanted experiences and distraction from values. The Spanish version of the VQ was adapted following the guidelines of the International Test Commission for cross-cultural instrument adaptation.

Other outcome measures

The following complementary instruments were used to assess convergent and divergent validity:

1. Acceptance and Action Questionnaire–II (AAQ-II): A 7-item instrument with a 7-point Likert-type scale measuring experiential avoidance. The Spanish version showed excellent internal consistency (α = 0.90) and a one-factor structure.

2. Cognitive Fusion Questionnaire (CFQ): A 7-item questionnaire with a 7-point Likert-type scale measuring cognitive fusion. The Spanish version showed internal consistency of α = 0.92 and a one-factor structure.

3. Mindful Attention Awareness Scale (MAAS): A 15-item scale with a 6-point Likert-type scale evaluating awareness of the present moment. The Spanish version exhibited a one-factor structure and internal consistency of α = 0.92.

4. Satisfaction with Life Scale (SWLS): A 5-item questionnaire with a 7-point Likert-type scale assessing subjective perception of quality of life. The Spanish version showed a one-factor structure with α = 0.89.

5. General Health Questionnaire–12 (GHQ-12): A 12-item instrument with a 4-point Likert-type scale assessing psychological difficulties. The Spanish version showed internal consistency of α = 0.90.

6. Depression, Anxiety, and Stress Scales–21 (DASS-21): A 21-item questionnaire with a 4-point Likert-type scale evaluating three negative emotional states (depression, anxiety, stress). The Spanish version showed a second-order factor structure with good internal consistency for all subscales.

Data analysis

The following statistical analyses were conducted:

1. Correlation analysis: Corrected item-total correlations were calculated for each VQ subscale using SPSS 25.

2. Confirmatory factor analysis (CFA): CFAs were conducted using LISREL 8.71 with robust maximum likelihood (MLR) estimation due to lack of multivariate normality in the data (Mardia's test: skewness = 2103.08, p < .001). Both a two-factor model and a two-factor model with correlated error terms for Items 5 and 7 were evaluated.

3. Goodness-of-fit indices: Five indices were computed: root mean square error of approximation (RMSEA), comparative fit index (CFI), non-normed fit index (NNFI), expected cross-validation index (ECVI), and standardized root mean square residual (SRMR). RMSEA < 0.08 was considered good fit, CFI and NNFI > 0.90 acceptable fit and > 0.95 good fit, and SRMR < 0.08 good fit.

4. Measurement invariance: Measurement invariance across gender and samples (clinical vs. nonclinical) was evaluated through multigroup analysis. The baseline model, metric invariance, scalar invariance, and strict invariance were estimated. Differences in RMSEA (ΔRMSEA) < .01, CFI (ΔCFI), and NNFI (ΔNNFI) < .01 were used as criteria to evaluate invariance.

5. Latent mean differences: Given that measurement invariance was established, latent mean differences were compared across gender and clinical status groups.

6. Convergent validity: Pearson correlations were calculated between VQ subscales and other relevant psychological measures. Cohen's criteria were used: small correlation (r between 0.10 and 0.20), medium correlation (r between 0.21 and 0.36), and strong correlation (r > 0.37).

Results

Psychometric quality of items

VQ items demonstrated adequate discrimination indices. For the Progress subscale, corrected item-total correlations ranged from 0.44 to 0.68 in Sample 1, 0.48 to 0.75 in Sample 2, and 0.57 to 0.71 in Sample 3. For the Obstruction subscale, correlations ranged from 0.53 to 0.59 in Sample 1, 0.52 to 0.70 in Sample 2, and 0.52 to 0.71 in Sample 3. Overall, results suggest that all VQ items demonstrated at least an adequate level of discrimination.

Internal consistency

Cronbach's alphas for the Progress subscale were: Sample 1 = .81 (95% CI [.79, .83]), Sample 2 = .85 (95% CI [.83, .86]), Sample 3 = .85 (95% CI [.82, .87]), with an overall alpha of .83. For the Obstruction subscale, alphas were: Sample 1 = .78 (95% CI [.75, .80]), Sample 2 = .84 (95% CI [.83, .86]), Sample 3 = .83 (95% CI [.80, .86]), with an overall alpha of .82. These values indicate adequate to good internal consistency.

Confirmatory factor analysis

The two-factor structure fitted the data adequately in the overall sample. The two-factor model obtained acceptable goodness-of-fit indices: RMSEA = 0.061 (90% CI [0.054, 0.068]), CFI = 0.98, NNFI = 0.97. The SRMR was 0.051. These indices remained relatively stable when error terms for Items 5 and 7 were allowed to correlate (RMSEA = 0.062, 90% CI [0.055, 0.069]; CFI = 0.98; NNFI = 0.97; SRMR = 0.051). However, the two-factor model without error term correlations was selected due to its greater parsimony.

The two-factor model also showed good fit for each sample separately:

• Sample 1: RMSEA = 0.058 (90% CI [0.047, 0.069]), CFI = 0.97, NNFI = 0.98, SRMR = 0.056
• Sample 2: RMSEA = 0.077 (90% CI [0.066, 0.088]), CFI = 0.98, NNFI = 0.98, SRMR = 0.059
• Sample 3: RMSEA = 0.063 (90% CI [0.045, 0.081]), CFI = 0.98, NNFI = 0.98, SRMR = 0.055

When nonclinical samples (Samples 1 and 2) were merged, the model also showed good fit (RMSEA = 0.062, 90% CI [0.055, 0.070], CFI = 0.98, NNFI = 0.97, SRMR = 0.053).

Measurement invariance

Measurement invariance across gender and samples (clinical vs. nonclinical combined) was evaluated. Changes in RMSEA, CFI, and NNFI were all below .01, indicating that criteria for metric, scalar, and strict invariance were met. Specifically:

• Metric invariance (factor loadings invariant across groups): ΔRMSEA = 0.006, ΔCFI = -0.001, ΔNNFI = 0.002
• Scalar invariance (factor intercepts invariant across groups): ΔRMSEA = 0.000, ΔCFI = -0.003, ΔNNFI = 0.000
• Strict invariance (error variances invariant across groups): ΔRMSEA = 0.003, ΔCFI = 0.000, ΔNNFI = 0.003

These findings indicate that the factorial structure of the VQ is equivalent across gender and clinical status.

Latent mean differences

Latent means of Progress and Obstruction were compared between gender and clinical status groups.

Regarding gender, females scored higher on Progress than males, but the difference was not statistically significant (ΔM = 0.10, SE = 0.05, p = .072). There were no statistically significant differences in Obstruction between genders (ΔM = -0.02, SE = 0.08, p = .820).

Regarding clinical status, the clinical group scored statistically significantly lower on Progress (ΔM = -0.39, SE = 0.06, p < .001) and significantly higher on Obstruction (ΔM = 0.85, SE = 0.10, p < .001) than the combined nonclinical sample. These findings are consistent with the expectation that individuals with clinical mental health difficulties experience lower progress in valued living and greater obstruction.

Convergent validity

Correlations between the VQ and other psychological measures were in the expected direction and magnitude, providing evidence of convergent validity.

Progress showed: medium to strong positive correlations with the MAAS (Mindfulness) and SWLS (Life Satisfaction) scales, medium to strong negative correlations with AAQ-II (Experiential Avoidance), CFQ (Cognitive Fusion), and DASS-21 (Depression, Anxiety, Stress), and negative correlations with GHQ-12 (Psychological Difficulties). Specifically, the average correlation of Progress with MAAS was r = 0.36, with SWLS was r = 0.56, with AAQ-II was r = -0.36, and with DASS-21 averaged r = -0.27 to -0.44.

Obstruction showed: the opposite pattern of correlations. Correlations of Obstruction with MAAS were moderate positive (average r = 0.29), with SWLS were negative (average r = -0.38), with AAQ-II were strong positive (average r = 0.62), with CFQ were strong positive (average r = 0.64), and with DASS-21 were strong positive (average r = 0.53 to 0.63).

Interestingly, correlations in Sample 2 (general population) were stronger than in the other samples, which may be due to greater heterogeneity and variability in scores.

Discussion and conclusions

This study has demonstrated that the Spanish version of the Valuing Questionnaire possesses adequate psychometric properties in Colombian clinical and nonclinical samples. Several findings warrant emphasis:

1. Factorial validity: The two-factor structure (Progress and Obstruction) remained stable and demonstrated good fit to the data across all samples analyzed. These results are consistent with prior research validating the VQ in English-speaking contexts.

2. Reliability: The internal consistency of the VQ was adequate, with Cronbach's alphas ranging from .78 to .85. These values exceed the recommended minimum of .70 and are similar to those reported in previous studies with English-speaking populations.

3. Measurement invariance: A particularly important finding was that the VQ meets criteria for metric, scalar, and strict invariance across gender and clinical status (clinical vs. nonclinical). This justifies making score comparisons between these groups and suggests that the observed differences in latent means (with the clinical sample showing lower Progress and higher Obstruction) reflect real differences in values rather than measurement biases.

4. Clinical differentiation: The VQ demonstrated capacity to differentiate between clinical and nonclinical samples, with the clinical sample obtaining significantly lower latent means on Progress (ΔM = -0.39, p < .001) and higher on Obstruction (ΔM = 0.85, p < .001). This finding underscores the utility of the instrument in clinical research.

5. Convergent validity: Correlations between the VQ and instruments measuring related constructs were in the expected directions and of appropriate magnitudes. Specifically, Progress correlated positively with measures of wellbeing (SWLS, MAAS) and negatively with measures of psychological difficulty (AAQ-II, CFQ, DASS-21, GHQ-12). Obstruction showed the inverse pattern. These findings provide strong evidence for convergent and discriminant validity.

6. Limitations and future directions: The study presents some limitations deserving consideration. First, the clinical sample was limited to emotional disorders (depression and anxiety), which limits generalization to other types of psychological difficulties such as psychosis or personality disorders. Second, data were collected using only self-reports, without formal diagnostic measures. Third, the mean age was relatively low across all samples, particularly in Samples 1 and 2, which limits generalization to older populations. Fourth, the majority of participants lived in Bogotá, so results may not generalize to other Colombian regions or other Spanish-speaking countries.

Despite these limitations, this study provides strong preliminary evidence that the Spanish version of the VQ is a psychometrically valid and reliable instrument for measuring valued living in Colombian clinical and nonclinical samples. Future studies should examine the psychometric properties of the VQ in more diverse samples, including older participants, individuals with different types of psychological difficulties, in other regions of Colombia, and in other Spanish-speaking countries.

Significance and contribution

This study contributes to the field of clinical psychology and psychological assessment by demonstrating that the Spanish version of the Valuing Questionnaire is a psychometrically valid and reliable instrument for measuring valued living in both clinical and nonclinical Colombian populations. The VQ maintained robust psychometric properties including clear two-factor structure, excellent internal consistency, measurement invariance across gender and clinical status, and theoretically coherent correlations with related constructs, supporting its use in research and clinical evaluation of valued living in Spanish-speaking contexts.

---