This concept refers to the belief and confidence in the own abilities to perform tasks involved in his or her job (Schyns & von Collani, 2002; Rigotti et al., 2008). People with high occupational task self-efficacy believe in their ability to move their cognitive resources and the course of action necessary to take control over the events of their lives (Bandura, 1977). They feel capable of performing tasks and perform them better. People with high occupational task efficacy persist in it, even in adversity, and they tend to interpret the problems more as challenges than threats. They increase their effort in the face of possible failures, and in case this happens they quickly recover their sense of effectiveness. Likewise, they face threats with the assurance that they can control those (Martínez & Salanova, 2006).

This factor involves “employee’s confidence in his/her ability to engage in the social interactional tasks necessary to initiate, maintain and develop interpersonal relationships at work” (Smith & Betz, 2000, p. 286). Employees with high social self-efficacy can complete effective social tasks at the workplace, as well as develop and maintain good relationships with others in the organization. They are probably going to like their co-workers and get help from them (Fan et al., 2013). In addition, social self-efficacy can be essential for building a team climate (Loeb et al., 2016)

Bandura (1997) defines emotional self-efficacy as an employee’s confidence in his or her ability to perceive, understand, regulate and use emotional information at work. Emotional self-efficacy includes a network of beliefs regarding one's ability to achieve results at different levels of specificity of emotional functioning (Schutte, 2018). Thus, these employees are able to recognize emotions and help to modify the emotions of others (Kirk, Schutte, & Hine, 2008, 2011). However, several studies have shown the importance of differentiating two dimensions: self- and other-oriented emotional self-efficacy (Choi, Kluemper, & Sauley, 2013; Dacre Pool & Qualter, 2012; Eklund, Loeb, Hansen, & Andersson-Wallin, 2012; Loeb et al., 2016). Self-oriented dimension focuses on personal development, and the other-oriented dimension aims to the relationships, interaction, and group dynamics (Loeb et al., 2016). Self-oriented emotional self-efficacy may be relevant for reducing emotional irritation and exhaustion (Fan et al., 2013; Loeb et al., 2016). Other-oriented emotional self-efficacy could be more relevant for leaders because they have to care and understand their staff (Deschênes, Dussault, & Fernet, 2016; Loeb, 2016; Loeb et al., 2016).

According to Cherniss (1993), professional self-efficacy is defined as people's beliefs in the ability to correctly fulfil their professional role. This construct involves workers’ beliefs about their own abilities to carry out successfully activities related to their profession (Maffei et al., 2012). These beliefs affect the choice and development of new activities. People tend to avoid activities that they think exceed their capabilities and perform those they consider capable of mastering. It has been observed that the beliefs of professional self-efficacy determine the amount of time that workers spend solving problems (Salanova, Bresó, & Schaufeli, 2005). Based on this concept we expect:

Hypothesis 1. Self-efficacy at work will be positive correlated with professional self-efficacy.

Several studies suggest that gender roles can affect self-efficacy scores. In terms of occupational self-efficacy, previous research shows that men report higher occupational self-efficacy than women (Abele & Spurk, 2009; Deschênes et al., 2016; Loeb et al., 2016; Williams & Betz, 1994). However, these differences vary according to the work domain: men have high scores in business and logical skills while women reported high self-efficacy in social and artistic domains (Feehan & Johnston, 1999; Koumoundourou, 2004; Lucas, Wanberg, & Zytowski, 1997).

Although the role of the female gender is associated with a social nature and it is expected that women report greater social self-efficacy than men (Eagly & Wood, 2011), Loeb et al. (2016) report no differences in gender for social self-efficacy. For emotional self-efficacy, previous research indicates that women are generally perceived to be better identifying and dealing with emotions of others, and understanding their own emotions (Deschênes et al., 2016; Loeb et al., 2016). Based on these previous studies, it is hypothesized that:

Hypothesis 2a. Men will report higher task self-efficacy than women.

Hypothesis 2b. Women will report a higher social self-efficacy than men.

Hypothesis 2c. Women will report higher self-oriented and other-oriented emotional self-efficacy than men.

Previous research has found that managers that have higher task self-efficacy are considered more reliable in their ability to complete work demands (Locke, Frederick, Lee, & Bobko, 1984). In addition, solidarity behavior leadership and positive feedback are effective to the development of other's self-efficacy (Bandura, 1997, 1999; Gist & Mitchell, 1992). Loeb et al. (2016) have found that social self-efficacy is an important predictor of team climate over and above occupational task self-efficacy. Luthans and Peterson (2002) reported positive correlations between self-efficacy and managers’ engagement with their work and ratings of managerial effectiveness. Furthermore, high self-efficacy has been found to relate to successful task leadership, higher levels of motivation, and the application of greater effort (Schyns & Sanders, 2005). Based on previous studies we hypothesized:

Hypothesis 3. Employees with people in charge will report higher self-efficacy, social and emotional self-efficacy than employees without people in charge.

A non-probabilistic sample of 771 employees was selected. 58% of the participants were female and 42% were male. Subjects ranged in age from 18 and 76 (M = 31.79; SD = 11.09). Regarding the level of education, 55% of the participants had an incomplete university education level, 24.8% had a complete university education level, 13.1% postgraduate level and 7.1% complete high school education level.

Employee’s affiliation institution was from private (n = 619) and public (n =152) organizations. Participants’ average experience in their current job was 5.28 years (SD = 6.78) and the overall job experience was 10.92 years (SD = 10.09). 65.2 % (n = 503) of employees reported no people in charge, the 23.1 % (n = 178) have between one and five, and 11.07% six or more people in charge.

Firstly, the English version was translated into Spanish. It was used a back-translation method by three independent translators taking into account the guidelines of the International Test Commission (ITC). The translators were bilingual in both English and Spanish languages, with experience in the content of the test, and with knowledge of assessment principles. Two self-efficacy experts evaluated the relevance, conceptual, and cultural equivalence of each item’s content using a three-option graduated scale (acceptable, not clear, not relevant). Subsequently, a pilot study was carried out whereby the selected items were administered to 10 workers of the Metropolitan Area of Buenos Aires (Argentina) over the age of 18.

Second, the employees were invited to participate in the research through social networks. Those interested in participating were contacted individually via email so that they could contact the experimenter if they wished. Each participant was informed of the objective of the investigation, the process and the commitment to confidentiality and anonymity of their participation. Participants recorded their voluntary participation agreement in an informed consent form, which also included the information that they could withdraw from the study at any point. The questionnaires were completed from February to September 2018. The instruments were administered online via Google Forms platform and the session lasted 20 to 30 minutes. There was no missing data because online forms used forced response options. The datasets analyses during the current study are available from the corresponding author on reasonable request.

The following descriptive measures were calculated for each item: mean, standard deviation, asymmetry, and kurtosis. The compliance with the statistical assumptions of univariate normality, linearity of the relationships, and absence of multicollinearity, were evaluated. An Exploratory Factor Analysis (EFA) was carried out with a subsample of 198 employees in order to explore the factor structure of the self-efficacy construct. Confirmatory Factor Analyses with a new sample (n = 470) were also carried out to test the model. Data analysis was performed using JASP-Stats 14.01 (The JASP Team, 2020). The goodness of fit was evaluated using the Chi-square, Standardized Root Mean Square Residual (SRMR), the Root Mean Square Error of Approximation (RMSEA) and the Comparative Fit Index (CFI). Total reliability and reliability of each factor as internal consistency was estimated using Cronbach's alpha coefficient.

Finally, Pearson correlations (n = 103) were calculated between factors of Self-efficacy at work scales and Professional Self-efficacy to provide evidence of convergent validity of the first scales. In addition, to analyze the concurrent validity, t-student tests (n = 771) were conducted for sociodemographic characteristics (gender: two categories; people in charge: two categories) and self-efficacy scales.

The study has been carried out considering the ethical principles of the Declaration of Helsinki (World Medical Association, 2013) and following the professional guidelines produced by American Psychological Association (APA, 2022). One of the principal concerns on online research is maintaining the anonymous and confidential information, protecting the vulnerability of participants. Therefore, the authors are reliant on participants' understanding and using privacy settings informed in the consent to prevent unintended sharing of personal information and research involvement. The consent was obtained electronically due to the risks for the subjects were low.

No atypical cases were identified. The analysis of multicollinearity shows that there were no inter-item correlations equal to or greater than .90 (Tabachnick & Fidell, 2001). In addition, all the asymmetry and kurtosis values are less than ± 2, within the normality parameters based on the criteria reported by George and Mallery (2019). However, Shapiro-Wilk Tests show non-normal distributions (p < .001).

The distributions Kaiser Meyer-Olkin (KMO) Measure of Sampling Adequacy was .79 indicating that the data were appropriate for EFA (Kaiser, 1974). Bartlett’s Test of Sphericity was significant (X² = 1206.649, p < .001). Based on these results, the matrix data were considered suitable for a factor analysis (EFA). Following the recommendations of Lloret-Segura, Ferreres-Traver, Hernández-Baeza and Tómas-Marco (2014), Weighted Least Square estimation method was used taking into account the ordinal data, non-normality distribution of the data, and the input for each analysis was the covariance matrix of the items. Parallel analysis was used as extraction method, suggesting a three-factor solution that explained the 38% of the variance of Self-efficacy at work. We adopted an oblique rotation method because there was a substantial theoretical basis to expect correlation among common factors (Fabrigar & Wegener, 2012; Tabachnick & Fidell, 2001). Three items loading under cut-off point .40 (Lloret-Segura et al., 2014) were removed due to the cross-loading criteria (items 4, 8, 15). Table 1 shows the matrix of factor loadings for the solution of factors. Based on the content of the items and the original dimensions, we decided to name the three factors as it follows: 1) Social Self-efficacy, 2) Occupational task Self-efficacy, and 3) Self-oriented Emotional Self-efficacy (See Table 1).

Table 1
Factor Loadings of Occupational Social, Task, and Emotional Self-efficacy scales.

Note. Extraction method is Weighted Least Square Mean. Applied rotation method is oblimin. *Item load under .40

Factorial Confirmatory Analyses (CFA) were conducted to test two different factor solutions (A) the original structure (four-factor solution) using all 19 items, and (B) three-factor solution without items 4, 8, 15, (version of 16 items). In all models’ solutions, the factors were allowed to correlate with one another and were analyzed using the estimator Diagonally Weighted Least Squares, Robust Method (Li, 2016). All fit statistics listed in Table 2.

Indicative criteria of very good fit as the Standardized Root Mean Square Residual (SRMR), the Root Mean Square Error of Approximation (RMSEA; with a confidence interval of 90%, and adequate values < .07), and the comparative fit index (CFI > .90) were used to analyze model fit and for comparison of the models (Hu & Bentler, 1999). In the table 2, the fit indices revealed that both models have shown good fit index. Therefore, the four-factors of the original scale with 19 items solution have shown a better fit to the observed data compared with the three-factor solution (Model B, 16 items). The four-factors model involved the following dimensions: 1) Occupational task Self-efficacy (6 items), 2) Social Self-efficacy (5 items), 3) Self-oriented Emotional Self-efficacy (4 items) and 4) Other-oriented Emotional Self-efficacy (See Figure 1).

The results indicate an appropriate internal consistency for overall self-efficacy at work (α = .85) and for each factor (Occupational task Self-efficacy α = .73; Social Self-efficacy α = .71; Self-oriented Emotional Self-efficacy α = .75; Other-oriented Emotional Self-efficacy α = .79).

Table 2
Fit Indices for the two models of the Self-Efficacy at work Scale

Note. RMSEA = Root mean square error of approximation; CFI = Comparative fit index

Values higher than 0.90 for CFI and RMSEA below 0.08 indicate good fit. The best fit model is in bold. Estimator: DWLS method Robust.

Figure 1
SEM Model C Standardized Solution.

Note. Fc1) Task Self-efficacy; Fc2) Social Self-efficacy; Fc3) Self-oriented Emotional Self-efficacy; Fc4) Other-oriented Emotional Self-efficacy.

Convergence studies consist of evaluating the extent to which correlations are observed between the scores of instruments that evaluate theoretically related or similar constructs. The scores obtained on the Self-efficacy at work scale were correlated with those obtained in the AU-10 questionnaire. As expected, all three dimensions were positively associated with professional self-efficacy. A strong correlation between “Total self-efficacy at work score” and “professional self-efficacy” was observed (r = .648; p < .001), which suggests that both instruments measure similar constructs. In addition, positive correlations were found between professional self-efficacy score and each dimension: Occupational task Self-efficacy (r = .607; p < .001), Social Self-efficacy (r = .423; p < .001), Self-orient Emotional Self-efficacy (r = .565; p < .001) and Other-oriented Emotional Self-efficacy (r = .358; p < .001) (See Table 3).

Table 3
Pearson Correlation between Self-efficacy at work and Professional Self-efficacy

Note. ** correlation is significant at .01 level (2 tailed)

Analysis of t-student tests was conducted for occupational task, social, and emotional self-efficacy. There were differences of gender on task self-efficacy men reported higher task self-efficacy (M = 4.19; SD = .56) than women (M = 4.08; SD = .60). There were no gender differences for social self-efficacy (t₍₇₆₉₎ = -1.30; p = .195; Cohen's d = -.096) and self-oriented emotional self-efficacy (t₍₇₆₉₎ = .260; p = .795; Cohen's d =.019). However, there is differences of gender on task self-efficacy (t₍₇₆₉₎ = 2.258; p = .024; Cohen's d =.165): women reported higher Other-oriented emotional self-efficacy (M = 3.94; SD = .74) than men (M = 3.82; SD = .75).

Results indicated differences between employees with/without people in charge. Employees with people in charge have more occupational task self-efficacy (M = 4.19; SD = .60) than employees without people in charge (M = 4.09; SD = .58; t₍₇₆₉₎ = -2.196; p = .028; Cohen's d = -.166). For social self-efficacy there was significant difference (t₍₇₆₉₎ = -2.284; p = .023; Cohen's d = -.173): employees with people in charge (M = 4.15; SD = .59) reported higher social self-efficacy than employees without people in charge (M = 4.05; SD= .62). No differences for self-oriented and other-oriented emotional self-efficacy dimensions were found.

One of the limitations is that a non-probabilistic sampling could produce some bias effects in the data set. Therefore, future studies should try to obtain a stratified random sample. In addition, self-efficacy is a construct that varies with age and experience; given that it ranges in age from 18 to 76 years old, these characteristics could have influenced the present results. Therefore, future research could consider a balanced age and experience group ratio.

The Spanish version of Self-efficacy at work Scales has adequate psychometric properties. Results demonstrated that this version has an adequate internal consistency, construct, and convergent validity to use in Argentinian employees. In addition, this adapted version maintains the original internal structure for Argentinian workers (four-factor model). Future studies should analyze the predictive validity of this instrument on job performance in different tasks according to their cognitive/personality demands. In addition, it could be interesting to study similar and different constructs related to outcomes at workplace for different professions. One promising area for future research is the integration of individual differences into existing models of performance. We agree with Judge, Jackson, Shaw, Scott and Rich (2007) that future research should analyze the joint effect of self-efficacy at work with personality traits on work behaviors.

Enviar correspondencia a: Robalino Guerra, P. E-mail: paulina.e.robalino@gmail.com