Participants

The study included 440 participants, all Colombian, with 321 women (72.95 %) and 118 men (26.82 %). Ages ranged from 18 to 81 years (M = 37.46; SD = 16.99). Data collection was conducted using a snowball sampling method. Table 1 presents information about the participants.

Table 1: Sample characteristics 

Instrument

The Beliefs About Psychological Services (BAPS) questionnaire was used. It consists of 18 items that measure individuals' beliefs about psychological services across three factors: intention (willingness or intention to seek psychological services -items 1, 2, 3, 4, 6, and 12), stigma tolerance (ability to cope with labels, stigma, and negative beliefs about psychotherapy -items 5, 8, 10, 11, 13, 15, and 17) and expertise (recognition and trust in the unique characteristics of professional psychological support-items 7, 9, 14, and 16). The questionnaire uses a Likert-type scale ranging from 1: strongly disagree to 6: strongly agree (^{Ægisdóttir & Gerstein, 2009}). In the Intention scale, a higher score indicates a greater willingness to seek or recommend psychological services; in the expertise dimension, higher scores reflect a positive evaluation of psychologists regarding their listening skills, support, and professional quality; in the stigma tolerance dimension, higher scores indicate a stronger stigma against seeking psychological services due to negative perceptions from others.

Procedure

The study followed the testing standards of the American Educational Research Association (^{AERA et al., 2018}). After obtaining approval from the original authors, the scale was translated into Spanish with the assistance of an expert in both English and Spanish (certified translator). Then, a back-translation into English was performed by another certified translator to compare it with the original version and ensure the equivalence of the terms. No discrepancies were found, confirming the translation's accuracy. Despite the existence of a Spanish version from Guatemala, a new translation was carried out based on testing standards to avoid cultural transformations introduced by prior translations (^{AERA et al., 2018}).

A pilot study was conducted with 30 participants via a Google Forms survey. They assessed the clarity and comprehension of the items and provided suggestions for improvement where needed.

Findings showed that most items were clear, though feedback was received for items 7, 16, 17, and 18. Minor wording adjustments were made without altering the meaning to improve clarity.

Once finalized, the survey was distributed digitally via social media and academic networks of the researchers and their institution to obtain a heterogeneous and representative sample. The online form included an invitation to participate, informed consent, a brief demographic questionnaire and the BAPS scale. A total of 459 participants responded. However, 19 were excluded for not meeting the criteria or refusing consent, leaving a final sample of 440 valid responses.

This study is part of the research project “Factores de riesgo y protección asociados a conductas de riesgo y problemas que afectan la salud mental en niños y adolescentes” (code: PSIPHD-4-2023), approved by the Research and Ethics Subcommittee of the Faculty of Psychology at Universidad de La Sabana.

Data Analysis

A confirmatory factor analysis (CFA) was conducted (due to the existence of a prior factor structure) using simulation models in EQS, applying maximum likelihood estimators and robust standardization to correct for non-normal multivariate effects (^{Lloret-Segura et al., 2014}). The expected fit indices were: Goodness of Fit Index (GFI), Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), Bentler-Bonett Non-Normed Fit Index (NNFI), Bentler-Bonett Normed Fit Index (NFI), Bollen’s Relative Fit Index (RFI), and Bollen’s Incremental Fit Index (IFI), with values above .90 considered acceptable, while error levels (Root Mean Square Error of Approximation-RMSEA, and Standardized Root Mean Square Residual-SRMR) were expected to be ≤ .08 (^{Hair et al., 2014}; ^{Samperio-Pacheco, 2019}).

Next, reliability analyses were conducted, including Cronbach’s alpha, McDonald's omega, Guttman’s λ6 indicator, and the Greatest Lower Bound coefficient, both overall and by factor, with expected values above .70. Multiple reliability statistics were used for greater accuracy-some (such as Cronbach’s alpha) are affected by correlation size, item count, or sample size, while others rely on variance, making them more rigorous. Including all of them enhances the robustness of the findings. Additionally, item deletion simulation was performed, expecting a decrease in coefficients as an indicator of each item's contribution to reliability, and item-total correlations were analyzed, expecting positive and significant values at p < .05, indicating each item's relationship with its dimension. Discriminant validity analyses were also conducted using Mann-Whitney U tests (due to non-normality) to assess the scale’s ability to differentiate between groups by comparing participants who attended psychological services versus those who did not. Finally, scoring norms were established for the Colombian context. Statistical analyses were performed using JASP software (version 17.1).

Translation, Back-Translation, and Pilot Testing

After the pilot testing, as evidence of content validity, it was found necessary to adjust formatting elements without altering equivalence with the original test. In item 7, the wording was changed to the first person, as third-person terms caused confusion among participants (original item 7: “Por su capacitación, los psicólogos lo pueden ayudar a encontrar solución a sus problemas”, adjusted: “Por su capacitación, los psicólogos pueden ayudarme a encontrar solución a mis problemas”). In items 16, 17, and 18, punctuation marks were added to improve the clarity of the statements.

Factorial Structure

Table 2 presents the confirmatory factor analyses as evidence of validity from the internal structure. Analyses were conducted with all items and then without item 18 (which showed low fit levels, low factor loading, and modification indices indicating that it loaded onto the other two factors).

Table 2: Confirmatory Factor Analyses 

As observed in Table 2, the goodness-of-fit indices show high values (above .90) and low error levels (below .08) in the test version without item 18. Table 3 presents the factor loadings, where it is evident that all items are explained by the factor to which they belong, with significant loadings (p < .001) and estimates above .50. The items in the stigma tolerance dimension also meet this condition, although their estimates are lower, which may be due to the fact that the items in this dimension are inverse to the others.

Table 3: Factor Loadings on the Scale Items 

Reliability Indicators

Concerning reliability, Table 4 shows the indicators for the general scale and by factor.

Table 4: Reliability Coefficients of the Scale 

When reviewing the data from Table 4, high reliability indices are evident, all above .75. It should be noted that reliability indicators based on variance rather than correlations, such as McDonald's ω, Guttman’s λ6, and the Greatest Lower Bound, are also high, reaffirming the reliability of the scale's dimensions and the BAPS overall. An analysis was conducted for each factor to simulate the removal of items, revealing that in all cases, the indicators were negatively affected, except for item 18. When item 18 was removed, the indicators improved, providing additional evidence of the reliability of the scale and each dimension. Finally, item-total correlations above .40 were also found, indicating an adequate internal consistency between the elements and their respective dimension.

Differences Between Those Who Attend or Do Not Attend Therapy

A Regarding discriminant validity, Table 5 presents comparisons of the dimensions and the total scale between individuals who attend therapy and those who do not.

Table 5: Comparisons Between Those Who Attend or Do Not Attend Therapy 

In Table 5, significant differences between the two groups can be observed, both in the intention and expertise dimensions, as well as in the overall score, indicating a statistically higher value for those who are attending therapy.

Scoring Norms

Percentile-based scoring norms are provided in Table 6 as evidence of validity through score interpretation, derived from the previously obtained differences.

Table 6: Scoring norms 

Study Limitations

A potential response bias related to selection bias is considered, as the sample did not exclude psychology students or professionals in psychological services, which could have influenced response transparency and trends, affecting internal validity. Additionally, 73% of the sample consisted of women, and previous research indicates that women, both nationally and internationally, tend to have a more positive attitude toward seeking psychological help compared to men, introducing a possible gender distribution bias (^{Hernández et al., 2014}; ^{Nizam & Nen, 2022}).

Similarly, socioeconomic strata distribution was unbalanced, with fewer participants from lower-income strata (1, 2, and 3), affecting sample diversity (^{Hernández et al., 2014}). Lastly, the snowball sampling method, being non-probabilistic and reliant on researchers' and participants' networks, limits full generalization of the findings despite the statistical robustness of the sample. It is recommended that future studies replicate this research using samples from multiple regions of the country with diverse social and cultural conditions to ensure broader applicability of the results.

The Colombian Ministry of Health and Social Protection (^{Minsalud & Colciencias, 2015}) highlights the need for studies identifying and addressing attitudinal barriers that may prevent individuals from seeking psychological services. Therefore, adapting the BAPS in Colombia is expected to generate knowledge that can inform interventions aimed at modifying beliefs about psychological services and increasing service utilization rates in the country. Future studies should further explore the hypothesis that self-stigma significantly influences decisions to seek psychological help. Investigating the specific relationship between these two variables would be valuable.

Additionally, predictive validity studies are recommended to determine whether test scores indicate a higher likelihood of future therapy attendance, along with convergent validity studies using similar scales or indicators. Studies comparing the scale’s metric invariance across demographic variables (e.g., place of origin, socioeconomic status) and examining belief differences based on these variables would be beneficial. This could lead to group profiling and the development of interventions that enhance and improve beliefs toward psychological services.