General self-efficacy (GSE) is a personal coping resource (Schwarzer, 1994) reflecting “one’s belief in one’s overall competence to effect requisite performances across a wide variety of achievement situations” (Eden, 2001, p. 75). GSE can also be described as “individuals’ perception of their ability to perform across a variety of different situations” (Judge et al., 1998, p. 170). Many studies show that expectations of competency have positive effects in different areas of life, such as health behavior and learning. Furthermore, it was shown that GSE is positively related to self-esteem, locus of control, and earning expectations (Bandura, 1997; Beierlein et al., 2013; Luszczynska et al., 2005). Because GSE affects the probability of success in many areas of life, there is great interest in evaluating this construct as a context variable for different research areas.

The FASKU complements the work of Beierlein and colleagues (2013) who developed the Allgemeine Selbstwirksamkeit Kurzskala (ASKU) based on the 10-item scale by Schwarzer and Jerusalem (1999), though the ASKU uses only three items to assess GSE. This scale is especially useful as it meets all scientific criteria for collecting and providing objective and valid data. Furthermore, given its features as a short scale, it is relatively easy to amend as a supplement to different types of surveys (Beierlein et al., 2013). The French version was developed and tested by Décieux et al. (2020). Based on the TRAPD approach (Translation, Review, Adjunction, Pretesting, and Documentation; Harkness, 2003), the items from the scale Allgemeine Selbstwirksamkeit Kurzskala (ASKU) were translated from the original German version by two independent translators who are both native French speakers. In the second step, the translated drafts were discussed and modified by a group of experts including translators, scientists, and other stakeholders. In a review process, the research team decided on one version. The final draft was evaluated using several steps implying a backtranslation and feedback from field staff and bilingual respondents before a final version was developed.

Samples

Data was collected in 2014 as part of a quantitative study of the Luxembourgish Youth Report 2015 (Schumacher et al., 2015). The study was implemented by the University of Luxembourg in collaboration with the Ministère de l’Education nationale, de l’Enfance et de la Jeunesse (MENJE) of Luxembourg and entailed a paper-and-pencil survey with 1,716 young people between 15 and 35 years. It consisted of 937 males (54.9%) and 770 females (45.1%).

Item analyses

The factor structure was tested with a confirmatory factor analysis (CFA; Décieux et al., 2020). As in the original study by Beierlein et al. (2013) the factor loadings of all three indicators were set to be equal (representing an essentially tau-equivalent measurement model; Graham, 2006). The MLR χ²-test statistic with robust standard error (Yuan & Bentler, 2000) was calculated and full-information maximum likelihood was used to account for missing data. Analysis was done with R version 3.6.0 (R Core Team, 2019) and the lavaan package (Rosseel, 2012). Results indicated that the single-factor model presented a good fit to the data for the total and the German version (Décieux et al., 2020). However, for the French version the RMSEA was quite high (see Table 2). Nevertheless, the confidence interval of the RMSEA covers a good model fit. Thus, the high RMSEA might also be due to a sample error. As the other fit indices were in an acceptable range, measurement invariance testing was conducted with the essentially tau-equivalent model as the basis for the configural model. The ΔCFI was used to assess goodness of fit of measurement invariance models. a CFI change of ≥ -.01 between a baseline model and the resulting model indicates measurement invariance (Little, 2013). According to the change in CFI, configural and metric invariance as well as scalar invariance were confirmed for the different language versions of the ASKU (see Table 3).

Item parameters

Table 4 details the results of the descriptive data analysis for the whole sample and the two different language versions of the ASKU.

Objectivity

Objectivity refers to the degree to which a measurement is independent of the examiner (cf. Jacob et al., 2019). This refers to different phases of an empirical study: the implementation, the evaluation and the interpretation. In the case of a PAPI survey, the implementation objectivity depends on the situation in which the respondent completes the questionnaire. Here, for example, the presence of third parties or other distractions may cause slight bias, but generally this should be a lesser problem than, for example, in interviews with interviewers, where this is only given if the interviewers follow the exact instructions and wording of the items when prescribing the scale. With appropriately trained interviewers, however, this is usually the case and ensures objectivity of implementation (Rammstedt, 2010). Evaluation objectivity concerns the numerical and categorical evaluation of respondents' answer behaviour according to fixed rules (cf. Lienert & Raatz, 1998). According to Beierlein et al. (2013), these are fully given for ASKU and thus also for FASKU, as the rules for calculating the values of the items are clearly defined and do not allow any room for interpretation. Interpretation Objectivity is given if the conclusions drawn from the survey results are comparable across different researchers. To maximise the objectivity of interpretation, researchers' knowledge of the measuring intention of the scale and of the interpretation of the quantitatively measured values should be comparable (Rammstedt, 2010). By standardising the evaluation and assigning a numerical measured value that describes respondents' level of general self-efficacy, the evaluation objectivity of the ASKU can also be considered as given according to Beierlein et al. (2013) and therefore also for the FASKU.

Reliability

Beierlein et al. (2013) already tested the reliability of ASKU. In our study (Décieux et al. 2020), reliability was satisfactory for the total sample (McDonald’s ω = .77) as well as for the two language versions (German: ω = .80; French: ω = .73).

Validity

Beierlein et al. (2013) showed that the ASKU is a valid measure of general self-efficacy. In our study, (Décieux et al., 2020) we additionally tested the construct validity of FASKU using intercorrelations to theoretically related constructs and corroborated the construct validity of the (F)ASKU scale (see Table 5 and Decieux et al., 2020). Moreover, we tested the original German Version of ASKU and the newly developed French Version (FASKU) for different forms of measurement invariance. According to the change in CFI, configural and metric invariance as well as scalar invariance were confirmed across the two language versions (see Table 3 and Decieux et al., 2020).

Descriptive Statistics

The overall mean of the ASKU was 3.86 (SD = 0.80). Men seem to be slightly more concerned with GSE (M = 3.91, SD = .81) than women (M = 3.80, SD = .79, t(1; 1,670) = 2.76, p = .001 d = −.14; 95% CI (−.23, −.04)). People who chose to complete the German language version of the ASKU had a mean score of 3.90 (SD = .77), and people who completed the French version (FASKU) had a mean score of 3.79 (SD = .84, t(1; 1,677) = 2.200, p = .01, d = −.14; 95% CI (−.23, −.04)).