The study sample comprised 2,007 individuals, with 132 students and 1,875 psychologists. The majority of participants were female residents of the Southeast region of Brazil, and nearly 60 % were married, with roughly half reporting having at least one child. Regarding education, which was assessed solely among psychologists, more than half held specialist qualifications, 20 % possessed master's degrees, and 5 % had completed doctorates. The average time since graduation was approximately 12 years.

Table 1: Note:

N: number of participants; χ²: Chi-squared test; p: significance level; Cramer’s V: effect size calculated to nominal variables; M: Mean; SD: standard deviation; η²: effect size calculated using the eta-squared test. * p< .05 *** p< .00

For participation in the study, the criteria we set required individuals to be at least 18 years old. We specifically sought fourth-year psychology students who were currently involved in clinical practice. Additionally, it was imperative for participants to have access to the internet and to be Brazilian citizens, a detail we confirmed through IP address checks.

On the other hand, we identified certain factors that would exclude potential participants. These included instances where the questionnaire submissions were incomplete or filled incorrectly. Additionally, while the survey was open, we intended to focus solely on psychology students and practicing psychologists. Therefore, responses from individuals from other professional backgrounds were excluded. Similarly, even if a response came from a Brazilian citizen, if it originated outside Brazil's geographic boundaries and this could not be cross-verified through IP addresses, we deemed it unsuitable for inclusion.

Participants initially provided sociodemographic information, including gender, region of residence in Brazil, marital status, number of children, and professional qualifications (among psychologists). Additionally, participants rated their perceived ability to handle various clinical demands, both in-person and virtually, on a scale ranging from 0 (totally insecure) to 10 (totally confident). This perception assessment was undertaken in alignment with existing psychometric validation studies employing perception as a component of convergent validity criteria (Andrade et al., 2020), particularly when alternative instruments are limited (Kline, 2015).

Psychologist and Counsellor Self-Efficacy Scale (PCES)

The PCES, originally developed in Australia (Watt et al., 2019), aims to assess psychosocial SE through 31 items categorized into five specific dimensions: (i) research; (ii) ethics; (iii) legal matters; (iv) assessment and measurement; and (v) intervention. The PCES was previously validated exclusively among Australian professionals, demonstrating high internal consistency (α = .94).

General Perceived Self-Efficacy Scale (GPSS)

The instrument assesses self-efficacy (SE) using ten items rated on a Likert scale, ranging from 1 to 4 points. The total score falls within a range of 10 to 40, with a higher score indicating a higher level of self-efficacy. Typical items include statements such as “Thanks to my resourcefulness, I know how to handle unforeseen situations” and “When confronted with a problem, I can usually find several solutions” (Schwarzer & Jerusalem, 1995). The original version of this instrument was developed in Germany by Schwarzer and Jerusalem in 1995 and has since been adapted for use in approximately 30 languages worldwide. It is widely recognized in the literature as one of the most frequently employed scales for assessing self-efficacy. Moreover, there is extensive evidence supporting its validity and reliability, drawn from research conducted with diverse populations across various countries (Scholz et al., 2002). The General Perceived Self-Efficacy Scale (GPSS) was adapted and validated for use in Brazil by Sbicigo et al. in 2012. This adaptation demonstrated good internal consistency, as indicated by a Cronbach's alpha coefficient of α = .85.

The data collection process extended over a six-month period, during which participants had access to the online questionnaire. On average, participants took about 25 minutes to complete the questionnaire comprehensively. The study adhered to the Declaration of Helsinki and received approval from the Research Ethics Committee of the Pontificia Universidade Católica de Campinas (protocol number 33780620.3.0000.5481).

The PCES was initially translated into Brazilian Portuguese by two bilingual psychologists. These translations were subsequently compared for semantics, concepts, linguistic nuances, item context, and idiomatic expressions. The primary researchers and translators collaborated to finalize the translation of items, adhering to specific cross-cultural adaptation and validation protocols outlined by the International Test Commission (ITC, 2005).

A pilot study, encompassing 10 students and 10 psychologists, was conducted to assess the conceptual, linguistic, and semantic equivalence of the PCES. This step facilitated participant comprehension of all PCES items. Any difficulties encountered in item comprehension were thoroughly discussed with the researchers, yet participants did not report any significant challenges in understanding the items.

The PCES was subsequently back-translated into English by two independent translators who were not involved in the initial translation phase, in accordance with specific guidelines (Wild et al., 2005). The synthesized back-translations were consolidated by the researchers and submitted to the PCES developers (Watt et al., 2019) for comparison with the original PCES version. Final Brazilian’ PCES version on the Appendix A.

The PCES was hosted on the SurveyMonkey^® platform, and the survey link was disseminated via various social networks and email. Participants could only complete the questionnaire once, with this restriction enforced based on IP addresses. Prior to commencing the questionnaire, all participants were required to read and consent to the Free and Informed Consent Form, accessible on the study website's homepage.

Sociodemographic data

Sociodemographic data were analyzed based on variable characteristics, with nominal variables assessed using the chi-square test (χ2) and continuous variables evaluated using one-way analysis of variance (one-way ANOVA). Effect size calculations were subsequently performed using Cramer's V or eta squared for nominal and continuous variables (De Almeida Lins et al., 2022; Lopes et al., 2022)

Internal consistency and factor structure

Internal consistency of the PCES was evaluated using Cronbach's alpha (α) and McDonald's omega (ω) coefficients. The utilization of both coefficients enabled performance comparison, ensuring higher analytical quality and reduced risk of bias in results. To assess factor structure, Confirmatory Factor Analysis (CFA) was conducted to determine if the original five factors of the PCES were retained (Watt et al., 2019). The Diagonal Weighted Least Squares (DWLS) estimator with robust error calculation was employed (Gaderman et al., 2012; Spritzer et al., 2022). Good model fit criteria included: Standardized Root Mean Square Residual (SRMR) less than .05; Tucker-Lewis Index (TLI) greater than .95; Root Mean Square Error of Approximation (RMSEA) less than .08; Comparative Fit Index (CFI) greater than .95; and a Chi-squared/degrees of freedom ratio (χ²/df) less than 3. These criteria were selected based on established guidelines (Cheung & Rensvold, 2002). To achieve the best-fit factor structure, all Modification Indices (MI) exceeding 10 (MI > 10.00) were analyzed, with consideration given solely to item covariances within the same dimension (Andrade et al., 2023a; Andrade et al., 2023b).

Following the establishment of the optimal factor structure (Model 3), a Multigroup Confirmatory Factor Analysis (MGCFA) was executed to assess measurement invariance across gender (male (n= 450) and female (n= 1,557)), participant profile (psychology students (n= 132) and psychologists (n= 1,887)), and age. Participants' ages were categorized as per guidelines from the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatística, 2020): (i) young adults (18 to 29 years, n = 262); (ii) adults (30 to 59 years, n = 1,484); and (iii) elderly (60 years or older, n = 261). Measurement invariance was considered present if differences in RMSEA and CFI values (Δ RMSEA and ΔCFI) were less than .02 and .01, respectively (Cheung & Rensvold, 2002).

Network Analysis (NA) was also conducted to estimate the clustering between PCES items (nodes) through partial correlations. NA enabled the identification of item grouping patterns and comparison with data derived from CFA and MGCFA. Key data from the 31 PCES items provided by NA included those with the highest number of connections with other items (betweenness centrality), the strength of these connections (degree centrality), the distance between nodes (closeness centrality), and variables indicative of the most influential factors in the network (expected influence) (Epskamp et al., 2012). LASSO (Least Absolute Shrinkage and Selection Operator) was used to construct graphs (Figures 1A and 1B) due to its ability to estimate correlations between PCES items (nodes) through partial correlations. This approach mitigates overfitting by excluding low-magnitude correlations (De Oliveira Pinheiro et al., 2022). In the graphical representation, nodes are depicted in green (positive correlation) or red (negative correlation), with line thickness indicating the strength of correlations between items. Item positioning within the graph's center is determined by the intensity and quantity of associations with other PCES items.

Convergent validity

Convergent validity was assessed through Spearman correlation analysis, examining the total PCES score in relation to PGSES and participants' subjective perception of their confidence in handling various specific demands typical of psychologists (as described in the instruments section).

As depicted in Table 1, significant differences in PCES total scores and four dimensions were identified between students and psychology professionals, with the exception of the assessment and measurement dimension. Gender, however, did not yield significant differences in the total PCES score and all its dimensions.

Regarding the factor structure of PCES (Table 2), the initial model indicated adequate factor adjustment based on the considered fit indices, confirming the instrument's original structure (Watt et al., 2019). Nevertheless, the factor structure was improved by incorporating modification indices aligned with the scale structure. Consequently, new modification index results were integrated into the second model, culminating in the final structure (Model 3), which exhibited superior fit indices compared to the previous models. Additionally, the χ²/gl ratio decreased, further enhancing the overall fit. The Multigroup Confirmatory Factor Analysis (MGCFA) also indicated favorable model fits when comparing the factor structure of PCES across gender, age groups, and professional profiles. Consequently, the identified levels of invariance implied that the instrument is suitable for use in various populations.

Table 2: Note:

CFA: Confirmatory Factor Analysis; RMSEA: Root-mean-square error of approximation; TLI: Tucker-Lewis Index: SRMR: Standardized root-mean-square residual; CFI: Comparative Fit Index; MGCFA: Multigroup Confirmatory Factor Analysis.

Table 3 presents the factor loadings of items categorized by their respective factors and the reliability indices of the PCES scale. In general, the data indicated a high level of instrument reliability, as determined through Cronbach's α (α = .951) and McDonald's ω (ω = .952) analyses. Notably, the intervention factor exhibited the highest reliability indices, while the research factor displayed the lowest indices.

Table 3: Note:

M: mean; SD: standard deviation. The overall Cronbach's α was .951 and the overall McDonald's ω was .952.

Figure 1 displays Gaussian graphical models for the five factors of PCES among psychology students (Figure 1A) and psychologists (Figure 1B). No significant differences were detected in the network structures between these two participant groups, as the distribution of items appeared similar. Notably, the legal matters dimension exhibited robust correlations among item pairs in both groups. Moreover, among psychologists, a strong correlation was observed between items 28 and 29. Additionally, item 17 displayed the highest centrality among students, whereas among psychologists, item 8 held the highest centrality (Figure 1C).

Figure 1:

Regarding the correlation levels with the PCES instrument (Table 4), the total instrument score exhibited positive correlations with all the presented variables. Notably, the intervention factor displayed the highest correlation strength among all the scale factors. A weak correlation was observed between the Perceived General Self-Efficacy Scale (PGSES) and the total PCES score. Furthermore, the total PCES score exhibited moderate correlations with certain therapist skills, particularly in face-to-face care situations. When assessing correlation indices between factors, the intervention factor demonstrated the highest frequency of moderate correlations with other evaluated items.

Table 4: Note:

All correlations presented in this table are statistically significant.