To answer this research question, multiple regression analyses were conducted to discover the impacts of antecedents on teachers’ self-efficacy and their pedagogic strategies. Accordingly, six models were set up. They are efficacy as a whole (Model 1: TSE), its three subscales, i.e., efficacy for student engagement (Model 2: EFSE), efficacy for classroom management (Model 3: EFCM) and efficacy for instructional strategies (Model 4: EFIS), CLT approach (Model 5: CLT) and non-CLT approach (Model 6: NCLT), as summarized in Table 4-2-1.
To sum up, multiple regression analyses manifested that antecedents
significantly contributed to the prediction of teachers’ self-efficacy, the prediction of its subscales, and the prediction of teachers’ pedagogic orientation toward CLT.
However, the antecedents in this study failed to explain teachers’ pedagogic orientation away from CLT. In addition, not all of the antecedents have significant effects on teachers’ self-efficacy. Their predictive effects are conditional upon the components of self-efficacy. Among the 13 antecedents, teachers’ satisfaction of their past teaching experience exerted the strongest influence on teachers’
self-efficacy.
In Model 1, using teachers’ self-efficacy as dependent variable and the 13 antecedents as independent variables, a multiple regression analysis was employed.
The result revealed that the 13 antecedents significantly contributed to the prediction of teachers’ self-efficacy (F=8.821, p<0.001) and explained 30.6% (R2=.306) of the total variance to teachers’ self-efficacy. That is, these 13 factors can predict teachers’
self-efficacy to some degree. In addition, a further comparison among these 13 antecedents displayed that teachers’ satisfaction of their past teaching experience (β
59
=.344, p<0.001) and students’ motivation of learning English (β=.114, p<0.05) demonstrated significant positive effects on teachers’ self-efficacy. That is to say, teachers who regarded their past teaching experience as a success would usually have higher self-efficacy, which means they believe they would have greater
influence in classroom management, student engagement and instructional strategies.
In addition, the teachers would also show stronger self-efficacy in a class where the students’ learning motivation was high.
Although the other remaining antecedents did not demonstrate a statistically significant impact on teachers’ self-efficacy, it is worth mentioning that whether the teachers had English class out of free will or not (β=-.081, p=0.104) and teachers’
self-evaluated pressure from weekly English periods (β=-.082, p=0.073) may more or less cause a negative influence on teachers’ self-efficacy. The data demonstrated that teachers’ self-efficacy was lessened when they brooded negative emotions over their teaching conditions.
In Model 2, using teachers’ efficacy for student engagement as dependent variable and the 13 antecedents as independent variables, another multiple regression analysis was executed to explore the effects of antecedents on this subscale (EFSE). The result exhibited that the 13 antecedents significantly foretold teachers’ efficacy for student engagement (F=8.502, p<0.001) and successfully answered 29.8% (R2=.298) of the variance to teachers’ efficacy for student engagement. The result showed that these 13 factors can predict teachers’ efficacy for student engagement to some degree. Furthermore, among these 13 antecedents, teachers’ satisfaction of their past teaching experience (β=.330, p<0.001) and students’ motivation of learning English (β=.138, p<0.05) demonstrated significant positive effects on teachers’ efficacy for student engagement. The findings clearly
60
exhibited that teachers gained confidence from their past successful teaching
experience and students’ positive feedback, which both helped to boost their efficacy to motivate and inspire students.
In Model 3, using teachers’ efficacy for classroom management as dependent variable and the 13 antecedents as independent variables, another multiple
regression analysis was executed to explore the effects of antecedents on this subscale (EFCM). The result displayed that the 13 antecedents significantly predicted teachers’ efficacy for classroom management (F=6.262, p<0.001) and successfully answered 23.7% (R2=.237) of the variance to teachers’ efficacy for classroom management. The result showed that these 13 factors have predictive power over teachers’ efficacy for classroom management to some degree. With further observation among these 13 antecedents, teachers’ satisfaction of their past teaching experience (β=.297, p<0.001) posed a significant predictive power over teachers’ efficacy for classroom management. Another background factor that also reached the significant level to predict teachers’ efficacy for classroom management is job position (β=.105, p<0.05). In the context of this research, job position consisted of three types—administrative staff, homeroom teacher, and pure subject teacher. This kind of job arrangement significantly helped the teachers to improve their competence in classroom management. In addition, even though students’
motivation of learning English(β=.096, p=0.075) did not show significant positive effects on teachers’ efficacy for classroom management, the impact was still
positively strong. One thing worth noticing is that teachers’ self-evaluated pressure from weekly English periods (β=-.130, p<0.01) put forth a significant negative effect on teachers’ efficacy for classroom management, which illustrated the stress teachers sensed has already damaged their competence in running a class effectively
61
and would probably make it difficult for them to deal with the disruptive behaviors in class.
In Model 4, using teachers’ efficacy for instructional strategies as dependent variable and the 13 antecedents as independent variables, another multiple regression analysis was executed to explore the effects of antecedents on this subscale (EFIS). The result displayed that the 13 antecedents significantly predicted teachers’ efficacy for instructional strategies (F=8.473, p<0.001) and successfully answered 29.6% (R2=.296) of the variance to teachers’ efficacy for instructional strategies. The data showed that these 13 factors can predict teachers’ efficacy for instructional strategies to some degree. With further contrast among these 13 antecedents, teachers’ satisfaction of their past teaching experience (β=.304, p<0.001) succeeded in acting a significant positive effect on teachers’ efficacy for instructional strategies. Again, the successful occurrences in the past helped to increase one’s self-confidence, which encouraged one to refine the original methods or to adopt new strategies and, understandably, led to higher efficacy. Another background factor that helped to advance teachers’ efficacy for instructional strategies and reached its significant level is years of English teaching experience (β=.127, p<0.05). Under the context of this research, the results meant that those teachers with longer English teaching years were survivors from earlier harsh
teaching conditions and had developed more excellent skills in conjuring up creative ideas to instruct a language class. One thing worth paying attention to is
“willingness to teach English class” (β=-.132, p<0.01) that put forth a significant negative effect on teachers’ efficacy for instructional strategies. That is to say, the unhappiness EFL teachers felt about English periods arranged against their will had worn down teachers’ flexibility and made their teaching rigid.
62
In Model 5, using teachers’ pedagogic orientation toward CLT as dependent variable and the 13 antecedents as independent variables, another multiple
regression analysis was executed to explore the effects of antecedents on teachers’
pedagogic orientation toward CLT (CLT). The results displayed that the 13 antecedents significantly predicted teachers’ pedagogic orientation toward CLT (F=3.6, p<0.001) and successfully answered 15.1% (R2=.151) of the variance to teachers’ pedagogic orientation toward CLT. The result showed that these 13 factors can predict teachers’ proclivity toward their use of CLT approach to some level.
With further weighing among these 13 antecedents, students’ motivation of learning English (β=.121, p<0.05) and participants’ gender (β=.142, p<0.01) succeeded in acting a significant positive effect on teachers’ pedagogic choice toward CLT. The result showed that teachers were more willing to use CLT approach when students responded to this method enthusiastically and teachers’ gender was a decisive power over teachers’ employment of CLT approach. One thing worth paying attention to is whether teachers had English class out of their free will or not (β=-.180, p<0.001) exerted a significant negative effect on teachers’ pedagogic proclivity toward CLT.
This finding signified that the teachers would rather adopt other methods than CLT approach if teaching English was not their choice in the first place.
In Model 6, in order to reveal how the antecedents influenced teachers’
pedagogic inclination away from CLT, another multiple regression analysis was executed to explore the effects of antecedents on teachers’ pedagogic propensity away from CLT (NCLT), using teachers’ pedagogic strategies that are against CLT as dependent variable and the 13 antecedents as independent variables. The results displayed that the 13 antecedents did not have significant effects (F=1.91, p=0.014) on teachers’ pedagogic orientation away from CLT and only accounted for 8.7%
63
(R2=.087) of the variance to teachers’ pedagogic orientation against CLT. Since the results of this model could not fully explain the effects of the antecedents on
teachers’ pedagogic orientation away from CLT, they were not displayed in Table 4-2-1.
64
Table 4-2-1 Summary of Multiple Regression Analyses of the Effect of Antecedents
on Teachers’ Self-Efficacy and Pedagogic Strategies toward CLT
Antecedents Model 1
TSE
Satisfaction of Past Teaching TSE refers to “teachers’ self-efficacy.”
EFSE refers to “efficacy for student engagement”.
EFCM refers to “efficacy for classroom management”.
EFIS refers to “efficacy for instructional strategies”.
CLT refers to “communicative language teaching”.