نظریه حرفه ای شناختی اجتماعی برای پیش بینی منافع و اهداف انتخابی در آمار میان دانشجویان روانشناسی اسپانیایی

This study investigated the usefulness of social cognitive career theory—SCCT (Lent, Brown, and Hackett, 1994) in predicting interests and goals relating to statistics among psychology students. The participants were 1036 Spanish students who completed measurements of statistics-related mastery experiences, self-efficacy, outcome expectations, interests and goals/intentions. Structural equation modeling (including multi-group analysis) was used to test the fit of the hypothesized models to the data. Results indicated support for SCCT as a way to predict students' interests in statistics and their intentions of engaging in academic or professional activities where statistics is used. Collectively, the predictors accounted for 50% of the variance in interests and for 77% of the variance in goals. Implications both for future research on SCCT and for intervention in statistics education are discussed.

Lent et al., 1994 and Lent et al., 2000 outlined a comprehensive framework based on social cognitive theory (Bandura, 1986) to explain and predict the process involved in the development of vocational and academic interests, how vocational and academic choices are made, and how academic performance is achieved. Since then, numerous empirical studies in the United States have tested models derived for social cognitive career theory (SCCT) with university students, mainly in math and science-intensive fields and activities (e.g. Bishop and Bieschke, 1998, Ferry et al., 2000, Gainor and Lent, 1998, Lapan et al., 1996, Lent et al., 2001, Lent, Brown, Schmidt, et al., 2003, Lent et al., 2005 and Lent et al., 2008), as well as in other areas (e.g. Diegelman and Subich, 2001, Fouad et al., 2002 and Tang et al., 1999). As the findings generally support SCCT propositions (see Sheu et al., 2010), in recent years researchers have also been testing social cognitive hypotheses in different cultural contexts and educational systems, including Japan (Adachi, 2004), Italy (Lent, Brown, Nota, & Soresi, 2003), Portugal (Lent, Paixão, da Silva, & Leitão, 2010) and Turkey (Özyürek, 2005). According to Lent et al. (1994) the core of SCCT proposes that the individual's interests reflect the interaction between his or her self-efficacy beliefs and outcome expectations over time. Therefore the determination to engage in a particular activity or to produce a particular outcome (choice goals/intentions) can be explained as a result of interests and self-reference beliefs. In addition to postulating relationships among key theoretical constructs, SCCT incorporates the effects of individual differences and environmental influences on interest, choice and performance. An important aspect of SCCT is its domain-specific nature, as the social cognitive view of person and context differs from that of trait-oriented career theories, given that “SCCT is more concerned with relatively dynamic and situation-specific aspects of people (e.g. self-views, future expectations) and their environments” (Lent & Brown, 2006, p. 13). Because the theory deals with domain-specific aspects of human functioning, theorists consider that social cognitive constructs should be examined within specific domains (Bandura, 1986, Bandura, 2006 and Lent and Brown, 2006). Therefore, the agenda for SCCT research includes testing social cognitive hypotheses across subject-matter domains and examining whether the relationships among social cognitive constructs are similar (see e.g. Betz and Hackett, 2006 and Fouad et al., 2002). In this context, the aim of the present study was to test social cognitive career theory in the field of statistics for non-mathematical university students. To our knowledge, this domain has not been examined by SCCT researchers before. Specifically, this research examined the usefulness of core SCCT interest and choice models in predicting statistics-related interests and choice goals among Spanish psychology undergraduates. Therefore, this study also seeks to extend previous research by testing SCCT hypotheses in a Spanish university setting. In addition, university students' difficulties in attaining an adequate performance in their statistics courses, coupled with their negative attitudes and beliefs about statistics, have been widely documented all over the world for over two decades (for a review see e.g. Blanco, 2008). Based on over 15 years' research and applied experience, SCCT can be viewed as a potential basis for conceptualizing and designing career-development interventions in the field of statistics, including efforts to expand interests and choice goals among students of social science, psychology, education and other non-mathematical subjects. This study examined several basic SCCT predictions about self-efficacy, outcome expectations, interests and goals. Specifically, the research focused on testing five principal hypotheses: Hypothesis 1. Statistics self-efficacy and outcome expectations predict interests in statistics. Hypothesis 2. Interests in statistics are predictive of the goals or intentions of students to engage in statistics-related activities in their future career. Hypothesis 3. Statistics self-efficacy contributes to statistics outcome expectations and also affects interests in statistics indirectly through outcome expectations. Hypothesis 4. Outcome expectations affect statistics-related intentions both indirectly (through interests) and directly. Hypothesis 5. Statistics self-efficacy has a significant indirect effect on goal/intentions through outcome expectations and interests. Since these five hypotheses concerning the SCCT core have received the strongest empirical support to date–and particularly those referring to direct effects– they were considered the principal predictions in this preliminary study. In addition, the study examined the role of mastery experiences in the development of statistics-related interests and intentions using an exploratory approach, by comparing two alternative models, model A and model B (see Fig. 1).According to social cognitive theory, self-efficacy is mainly developed through successful mastery experiences (Bandura, 1986). In prior research on SCCT, several studies tested the direct effect of mastery experiences on self-efficacy, and their results supported this proposition (Ferry et al., 2000, Fouad et al., 2002, Lapan et al., 1996 and Lopez et al., 1997). Therefore, the direct effect of personal accomplishment on statistics self-efficacy was included in both models (hypothesis 6a). As an exploratory extension of this hypothesis, it was hypothesized, also in both models, that mastery experiences have a significant indirect effect on the other factors (outcome expectations, interests and intentions) through self-efficacy (hypothesis 6b). In view of the fact that only a limited number of studies have examined the direct and indirect effects of mastery experiences on the rest of the social cognitive constructs, the results of this work may contribute to increasing the knowledge of the role played by the main source of self-efficacy in SCCT interest and choice models. Furthermore, one of the frequent findings of research into statistics education is the significant relation between performance or previous experience (constructs which are close to mastery experiences) and attitudes to statistics, the measure of which usually includes behavioral intentions (see Blanco, 2008). Therefore a tentative hypothesis was formulated in order to examine more closely the nature of the relationship between mastery experiences and intentions or goals. In model A the effect of mastery experiences on intentions was fully mediated by self-efficacy, so additional paths were not included. However, model B was defined as a partially mediated model in which mastery experiences affected goals/intentions both indirectly through self-efficacy and directly too (hypothesis 7). As shown in Fig. 1, model A and model B are different, as only the second includes the path mastery experiences→goals, that is, hypothesis 7. It was considered that all the predictions in this study, as well as the SCCT itself, could be generally applied to the reference population. Therefore the explanatory capacity of the selected model was assumed to be independent of other general variables of an academic nature, such as institutional context or time of assessment. In particular, the following invariance hypotheses were tested: Hypothesis 8. The relations in the model do not vary according to the university where the students are studying. Hypothesis 9. Assuming that the reference population has some prior experience in the field of statistics, the relations in the model do not vary according to the precise moment on the students' educational path.

Taken as a whole, the results of this study support the application of SCCT to explain the statistics-related interests and choice goals of Spanish psychology students. According to the interest model of Lent et al. (1994), statistics self-efficacy and outcome expectations directly predicted statistics interests. Statistics self-efficacy also appeared to influence statistics interests through outcome expectations. These results support past findings from SCCT models, and suggest that psychology students are more likely to form positive statistics-related outcomes and desires, and have interests in statistics when they believe in their statistical abilities. The results also indicate that students must not only believe that they are capable of performing statistics tasks but also understand the rewards of engaging in statistical activities. The choice model of Lent et al. (1994) also appeared to be at least partially applicable to the statistics field (it must be noted that the model tested in this study did not hypothesize that self-efficacy affects choice intentions directly). SCCT posits that people's intentions of engaging in statistics-related activities are related to their self-efficacy, outcome expectations, and interests. The findings of this study support these predictions. Self-efficacy produced a significant indirect path to intentions through interests and outcome expectations. In addition, outcomes expectations produced both a significant direct path to intentions and an indirect path to intentions through interests. Interests also directly affected choice intentions, with outcome expectations producing an even greater positive influence on intentions. Perhaps one of the most noteworthy findings in the present study was the empirical support for the importance of statistics-related outcome expectations in explaining choice goals or students' intentions of engaging in statistics-related activities in their future career. Social cognitive theory identifies mastery experience as the most influential source of self-efficacy (Bandura, 1986). Consistent with this proposition, the study also found evidence of a significant direct effect of mastery experience on statistics self-efficacy. Prior personal performance accomplishment in quantitative fields (mathematics and statistics) accounted for a substantial amount of the variance of current statistics self-efficacy. The findings also suggest that mastery experiences affect intentions both directly and indirectly through self-efficacy (partially mediated model B fitted the data better than fully mediated model A). These results do not concur with the proposals of SCCT, which define a model of effects that is completely mediated by self-efficacy. Collectively, the predictors in the model (mastery experience, self-efficacy and outcome expectations) accounted for 50% of the variance in interests and for 77% of the variance in intentions for the full sample. These proportions of explained variance are substantial when compared with the results of past research on SCCT in math and science-intensive fields (see Lent et al., 2008) as well as in other areas (see Sheu et al., 2010). The findings also indicate that the predictive usefulness of social cognitive variables is not moderated by university or by year of the degree. That is, SCCT variables may help explain statistics-related interests and intentions of students from different institutional contexts and among students at different times in their education. For a correct interpretation of the results of the present study it is necessary to consider its limitations. Probably the primary limitation of this research is the use of new measurements of statistics-related mastery experiences, self-efficacy, outcome expectations, interests and goals, without well known psychometric properties (i.e. validity and reliability), and containing relatively few items. This is especially true in the case of the mastery experiences construct, as it was based only on two items. Although the preliminary analyses and the goodness-of-fit indicators for the measurement model appear to be promising, new studies are clearly needed both to validate the instruments used in this study and to develop new scales (e.g. mastery experiences). This research must therefore be viewed as a preliminary approach to providing an initial insight on the SCCT model test in a Spanish university setting and in a specific domain (i.e., statistics). Another limitation of this work stems from the fact that cross-validation procedures were not used to specifically examine the hypothesis which was not derived from SCCT, the one formulated regarding mastery experiences (hypothesis 7). This analytical approach would make it possible to better assess whether the findings obtained in this study can be attributed to the specific characteristics of the sample used. It is therefore necessary to use a new sample for this purpose in the future. Meanwhile, the role of statistics-related mastery experiences in the model, and specifically the direct effect on intentions/goals, should be considered as a tentative finding. In the third place, one potential limitation of this work concerns the generalization of the results obtained in different student populations. The selection of the sample using non-random procedures may have introduced biases (for example students with a greater tendency to absenteeism may have been excluded from the sample). Moreover, as all the students in this study were Spanish psychology students, the application of the results to university students in other educational systems or in other specialties should be undertaken with caution. Future research would benefit from including participants from multiple majors, and additional studies of SCCT within a cross-national context would also permit a fuller appreciation of the generalizability of SCCT predictions as applied to statistics education in non-mathematics students. Finally, another potential limitation of the present study is the very low variance in statistics-related outcome expectations predicted by the model. That is, the positive and significant influence of statistics self-efficacy did not explain a substantial portion of variance in outcome expectations. Given that this study is the first time SCCT has been assessed in the context of the Spanish university system, it was decided initially to set up a simple model. However, future research in this field needs to examine the role of mastery experiences in shaping outcome expectations, and not only their impact on self-efficacy. It would also be useful to broaden the research to other learning experiences previously defined in social cognitive theory (vicarious experience, social persuasion and physiological and emotional states). Although the design of this present study limits the determination of casual relationships and order among variables in the SCCT model, some practical implications can be derived from the results. Two particular targets for intervention which can be derived from SCCT are the need to strengthen statistics self-efficacy beliefs, and to instill positive and realistic statistics-related outcome expectations. As both factors are essential prior requirements for interests in statistics-related activities and for students' intention of becoming involved with these, their reinforcement would go far towards improving outcomes in statistics education. Although more theory-based intervention research is needed (Lent et al., 2008) some researchers have already had some success in creating SCCT interventions as applied to general vocational areas (Betz and Schifano, 2000 and Diegelman and Subich, 2001). It would therefore be useful to design specific intervention strategies to raise statistics self-efficacy and to change the outcome expectations of non-mathematical students. In summary, the present findings support the use of SCCT to explain and predict the process involved in the development of statistics-related interests and choice goals in Spanish psychology students. These results contribute to SCCT research by extending empirical evidence about interest and choice models across domains and cultural settings. Furthermore, since one of the strengths of SCCT is that it is linked to interventions, the results of this study offer insights as to how to improve outcomes in statistics education, a challenge for many educators and students in non-mathematical majors around the world.