Validity Evidence of the Spanish Version of the Automatic Thoughts Questionnaire–8 in Colombia

Full reference: Ruiz, F. J., Suárez-Falcón, J. C., & Riaño-Hernández, D. (2017). Validity evidence of the Spanish version of the Automatic Thoughts Questionnaire–8 in Colombia. The Spanish Journal of Psychology, 20, e12. https://doi.org/10.1017/sjp.2017.7

Study type: Cross-cultural validation and factor structure analysis

Background and objectives

The Automatic Thoughts Questionnaire (ATQ) is one of the most widely used instruments in research and clinical practice for assessing the frequency of negative automatic thoughts in depressed individuals. Originally developed by Hollon and Kendall in 1980, the ATQ comprises 30 items rated on a 5-point Likert scale. Although the ATQ has consistently demonstrated excellent psychometric properties across multiple languages and cultural contexts, its length can constitute a practical obstacle in research requiring administration of comprehensive instrument batteries. This measurement efficiency problem can result in respondent fatigue and increased nonresponse rates. For this reason, Netemeyer et al. (2002) proposed abbreviated versions of the ATQ: a 15-item version (ATQ-15) and an 8-item version (ATQ-8). Subsequent studies confirmed that the ATQ-8 maintained a unidimensional structure and psychometric properties comparable to the full version, with Cronbach alphas between .85 and .92.

However, at the time of this study, no systematic validation of the ATQ-8 in Spanish version existed in Colombian samples. This gap is significant because validating instruments in culturally diverse samples strengthens both confidence in the measure and the cross-cultural relevance of the underlying theory. Additionally, Colombia lacked validated measures of negative automatic thought frequency, making it difficult to conduct studies on the cognitive model of depression in that context.

Method

Participants

The study included four samples with a total of 1,587 participants:

Sample 1 (n = 762): University students from seven universities in Bogotá, ages 18-63 (M = 21.16, SD = 3.76). Forty-six percent were studying Psychology, with the remainder in diverse disciplines (Law, Engineering, Philosophy, Communication, Business, Medicine, Theology). Sixty-two percent were women. Twenty-six percent reported prior psychological or psychiatric treatment, but only 4.3% were currently in treatment. 2.9% were taking psychotropic medication.

Sample 2 (n = 724): General Colombian population recruited through anonymous internet survey distributed via social media. Ages 18-88 (M = 26.11, SD = 8.93); 74.4% female. Educational level: 17.8% primary/secondary education, 63.8% undergraduate/bachelor, 18.4% postgraduate. 45% reported prior treatment, 8.4% were currently in treatment, and 5.4% were taking psychotropic medication.

Sample 3 (n = 101): Clinical sample of patients evaluated in private psychology consultation centers in Bogotá. Ages 18-67 (M = 32.22, SD = 12.09); 51.5% female. 67.3% presented emotional disorders and 32.7% sexual disorders. Only 5% were taking psychotropic medication.

Sample 4 (n = 11): Participants in a multiple baseline design study evaluating effects of a one-session Acceptance and Commitment Therapy (ACT) intervention focused on disrupting repetitive negative thinking. Ages 18-32 (M = 22.18, SD = 4.40); 81.8% female. All had been entangled in ruminative thinking for at least 6 months with significant interference in at least two life areas.

Instrument(s) under study

ATQ-8: Abbreviated version with 8 items on a 5-point Likert scale (5 = all the time, 1 = not at all). Items include statements such as "I'm no good," "What's wrong with me?," "I'm worthless," and "My future is bleak." The Spanish version of Cano-García and Rodríguez-Franco (2002) was used.

Dysfunctional Attitude Scale–Revised (DAS-R): 17-item measure assessing dysfunctional schemas across two factors: Perfectionism/Performance evaluation and Dependency (α = .91).

Depression, Anxiety, and Stress Scales–21 (DASS-21): 21-item scale measuring depressive, anxiety, and stress symptoms with hierarchical structure (three first-order factors and one second-order factor: Emotional Symptoms).

General Health Questionnaire–12 (GHQ-12): 12-item scale assessing general psychological distress (α: .88 to .93).

Satisfaction with Life Scale (SWLS): 5-item measure of self-perceived well-being (α: .84 to .89).

Acceptance and Action Questionnaire–II (AAQ-II): 7-item scale measuring psychological inflexibility or general experiential avoidance (mean α: .90).

Mindful Attention Awareness Scale (MAAS): 15-item measure assessing capacity to maintain present-moment attention versus operating on autopilot.

Data analysis

A small pilot study with 10 university students was conducted to confirm item comprehensibility for Colombian participants. All items were rated as highly comprehensible.

In Samples 1-3, questionnaires were administered with informed consent. Sample 1 completed them in classrooms during regular class time. Sample 2 completed an anonymous internet survey. Sample 3 responded during initial clinical assessment interviews.

In Sample 4, participants recruited through social media advertisements completed a baseline period of 2-10 weeks during which the ATQ-8 was administered every 2 weeks. They then received a one-session ACT intervention of approximately 75 minutes focused on identifying triggers for repetitive negative thinking, promoting acceptance and values clarification, and introducing cognitive defusion techniques. Follow-up measurements were conducted for 6 weeks (every 2 weeks).

Data were examined for missing values (0.60% total) that were imputed using similar response patterns in LISREL. Confirmatory factor analysis (CFA) was conducted using weighted least squares (WLS) estimation with polychoric correlations, appropriate for ordinal scales. Chi-square test and goodness-of-fit indices were evaluated: RMSEA, CFI, and NNFI.

Metric and scalar invariances were explored by comparing nested models using criteria from Cheung and Rensvold (2002): ΔRMSEA < .01 and ΔCFI/ΔNNFI ≥ -.01.

Cronbach alphas were calculated with 95% confidence intervals. Corrected item-total correlations, descriptive statistics, and gender differences were examined via t-tests. Pearson correlations were computed for convergent validity, and paired t-test was conducted to analyze treatment sensitivity in Sample 4.

Results

Item Quality and Internal Consistency

All items demonstrated good discrimination with corrected item-total correlations ranging from .46 to .78. Cronbach alpha ranged from .85 (Sample 1) to .91 (Sample 2), with overall alpha of .89 (95% CI [.88, .90]). No significant differences between men and women appeared in any sample.

Internal Structure

The one-factor model showed adequate fit: χ²(20) = 237.62, p < .01; RMSEA = .083, 90% CI [.074, .092]; CFI = .96; NNFI = .95. Fit values exceed acceptability criteria established by Hu and Bentler (1999), with CFI and NNFI > .95 indicating good fit. Standardized factor loadings ranged from .75 to .93.

Measurement Invariance

Metric and scalar invariance were supported across both the three samples and between men and women. Changes in RMSEA, CFI, and NNFI were less than .01, meeting invariance criteria. This finding is especially relevant because no previous study had explored measurement invariance of the ATQ-8.

Convergent Validity

The ATQ-8 showed theoretically coherent correlations with all assessed constructs:

• Significant positive correlations with: psychological inflexibility (AAQ-II: r between .56 and .71), depressive symptoms (r between .59 and .80), anxiety (r between .46 and .66), stress (r between .47 and .64), general emotional distress (GHQ-12: r between .50 and .60), and dysfunctional attitudes (DAS-R: r = .43).

• Significant negative correlations with: mindfulness/present-moment awareness (MAAS: r = -.29) and satisfaction with life (SWLS: r between -.44 and -.60).

All these correlation patterns aligned with predictions based on the cognitive theory of depression and the acceptance and commitment therapy model.

Criterion Validity

Comparative group analysis showed that mean ATQ-8 score in the clinical sample (M = 19.75, SD = 7.35) was significantly higher than in the student sample (M = 15.28, SD = 6.01; t = -5.75, p < .001) and general population sample (M = 17.41, SD = 7.46; t = -2.95, p = .004), demonstrating discriminative validity between groups.

Treatment Sensitivity

In Sample 4, mean baseline score was 21.45 (SD = 5.58) and decreased to 12.45 (SD = 4.78) at 6-week follow-up. The difference was statistically significant with very large effect size (t = 6.05, p < .001, Cohen's d = 1.61). This demonstrates that the ATQ-8 is sensitive to effects of the ACT intervention focused on disrupting repetitive negative thinking.

Discussion and conclusions

The Spanish version of the ATQ-8 demonstrated to be a valid, reliable, and culturally appropriate instrument for measuring negative automatic thought frequency in Colombian samples. The instrument maintains a clear unidimensional structure, excellent psychometric properties comparable to the original English version, established measurement invariance between demographic and clinical groups, and demonstrated sensitivity to therapeutic change. Authors identified several limitations: (1) ATQ-8 functioning was evaluated in a relatively small clinical sample, requiring future research in larger clinical samples; (2) systematic information on specific diagnoses was not obtained, with participants grouped in broad categories (emotional and sexual disorders); (3) some instruments used to explore convergent and discriminant validity lacked formal validation in Colombian samples, although their internal consistency in the present study was adequate.

Significance and contribution

This study contributes to the field of psychological assessment by providing robust validity evidence for an abbreviated measure of negative automatic thoughts in Spanish-speaking contexts. By demonstrating psychometric properties equivalent to the original version and establishing measurement invariance across demographic and clinical groups, the work expands the availability of validated instruments for research on the cognitive model of depression in Colombian and Latin American samples. Findings regarding treatment sensitivity suggest the ATQ-8 is a viable tool for assessing outcomes in clinical evaluation and therapeutic interventions in the region.