5.1.1 Self-perceived Proficiency & Metalinguistic Knowledge
There was a significant correlation between metalinguistic knowledge and self-perceived proficiency (r = 0.3). Besides, when metalinguistic knowledge was added to the hierarchical regression analysis with self-perceived proficiency as the dependent variable, metalinguistic knowledge did not exert a significant effect on self-perceived proficiency across the four Models when the effect of language achievement was controlled for and the motivational components and self-regulation were included into the analysis. However, metalinguistic knowledge held a significant predictive power on self-perceived proficiency when language achievement was not
66
controlled for (β = 0.3, p < .01).
It suggests that metalinguistic knowledge alone can predict self-perceived proficiency significantly. However, when the covariate (i.e. language achievement) was controlled for and other variables (e.g., motivational components and self-regulation) were included, its predictive power declined, contributing to the non-significant effect of metalinguistic knowledge on self-perceived proficiency.
When all the independent variables (i.e., metalinguistic knowledge, motivational components and self-regulation) are included altogether in the analysis, metalinguistic knowledge does not seem to play a significant role in influencing learners’
self-perceived proficiency. It is possible that motivational components and self-regulation will interfere with the effect of metalinguistic knowledge on self-perceived proficiency.
As previously mentioned, metalinguistic knowledge did not predict self-perceived proficiency significantly when the effects of metalinguistic knowledge, motivational components, and self-regulation were considered simultaneously in Model 4 of Table 4. Metalinguistic knowledge does not seem to be a significant indicator of learners’ perceptions of their proficiency, suggesting that sufficient metalinguistic knowledge cannot ensure learners’ positive evaluations of their proficiency, particularly when other factors are considered. It can be inferred from the findings that metalinguistic knowledge cannot predict learners’ perceptions of their proficiency and learners with sufficient metalinguistic knowledge of the target language may not necessarily have confidence in their own proficiency.
Besides, the non-significant effect of metalinguistic knowledge shown in Model 4 of Table 4 is partly in accordance with Perales and Cenoz’s (2002) study, in which the researchers found that participants’ self-evaluation of Basque proficiency did not correlate significantly with their metalinguistic awareness measured from Words in
67
Sentences of MLAT. Perales and Cenoz explained that learners with higher level of metalinguistic awareness might be too demanding about their proficiency to evaluate their proficiency highly and positively. However, the result reported by Perales and Cenoz (2002) can only support limitedly the findings discovered in the present study on account of the different instruments employed in both studies.
5.1.2 Self-perceived Proficiency & Self-regulation
According to the correlation analysis, self-perceived proficiency correlated significantly with self-regulation (r = 0.49). After adding self-regulation to the hierarchical regression analysis in Model 3, self-regulation exerted a significant effect (β = 0.49, p < .01) on self-perceived proficiency without controlling for the effect of the covariate (i.e. language achievement). Self-regulation still held a significant predictive power on self-perceived proficiency (β = 0.39, p < .01) when the effect of language achievement was controlled for and metalinguistic knowledge was included in the analysis. However, when metalinguistic knowledge, motivational components and self-regulation were included in the regression in Model 4, the effect of self-regulation not only reduced but became non-significant (β = 0.13).
When metalinguistic knowledge and self-regulation are juxtaposed for comparison, self-regulation is a stronger predictor of self-perceived proficiency since self-regulation had a much stronger significant effect on self-perceived proficiency in Model 3 while metalinguistic knowledge did not predict self-perceived proficiency significantly. However, when the motivational components were entered to Model 4, there was a great change in the predictive power of self-regulation—the predictive effect of self-regulation greatly declined and became non-significant. It is possible that the motivational components have a great impact on self-regulation, transforming self-regulation into a non-significant predictor of self-perceived proficiency.
68
Compared with the other variables in Model 4, self-regulation is not a significant predictor of self-perceived proficiency and it cannot predict or affect learners’
self-perceived proficiency. The degrees of learners’ self-regulation do not influence learners’ confidence in and perceptions of their proficiency. Self-regulated learners do not necessarily feel confident of their proficiency or evaluate their proficiency as highly as others do.
5.1.3 Self-perceived Proficiency & Motivation
The correlation analysis reveals that self-perceived proficiency had a significant correlation with motivation (r = 0.5). Self-perceived proficiency also correlated significantly with the five motivational components. When the motivational components were added to the hierarchical regression analysis with self-perceived proficiency as the dependent variable in Model 2, self-efficacy, extrinsic motivation and mastery orientation were found to be significant predictors of self-perceived proficiency no matter when the effect of language achievement was controlled for or not. Even when self-regulation was added to the analysis in Model 4, the above three components remained to be significant predictors of self-perceived proficiency and self-efficacy was still the best predictor of self-perceived proficiency (β = 0.47, p
< .01), followed by mastery orientation (β = 0.18, p < .01) and extrinsic motivation (β
= -0.16, p < .01).
According to Bandura (1986), self-efficacy refers to learners’ judgment of their abilities and competence to execute a course of action when participating in a task.
Learners with self-efficacy are more willing to participate in a task instead of avoiding it. Therefore, the above-mentioned findings suggest that learners who present a high level of self-efficacy are more likely to generate positive evaluations of their own proficiency since they are more willing to devote themselves to achieving tasks and
69
they have more confidence in their capabilities to accomplish the task. Mastery orientation also had a significantly positive effect on self-perceived proficiency and was a strong indicator of self-perceived proficiency. Learners learning a target language with the purpose of mastering the language and fulfilling their own goals and expectations are very likely to have positive evaluations of their proficiency. On the contrary, extrinsic motivation held a negative predictive power on self-perceived proficiency, suggesting that learners learning a target language out of extrinsic motivation may not evaluate their own proficiency positively. If learners learn a language in order to avoid punishment or to obtain awards, they may not have confidence in their abilities since extrinsically-motivated learners usually do not learn for their enjoyment of the language but for other external outcomes.
In addition, the above findings are similar to those discovered by Perales and Cenoz (2002), who found that instrumental orientation correlated with self-evaluation negatively and it presented a negative influence on self-evaluation. Perales and Cenoz (2002) suggested that instrumentally-oriented learners who learned Basque for job requirement may be more likely to have lower evaluations of their proficiency in Basque. The explanation is partly consistent with the finding in the present study that extrinsic motivation exerted an adverse effect on self-perceived proficiency and extrinsically-motivated learners who learn a language for external factors are likely to have lower evaluations of their proficiency.
In conclusion, self-perceived proficiency can be predicted by self-efficacy, extrinsic motivation and mastery orientation. Learners who learn a language with high self-efficacy or out of mastery orientation are more likely to have confidence in their proficiency whereas extrinsically-motivated learners would have negative evaluations of their proficiency since they mostly learn a target language for external factors such as awards or punishment avoidance. It is thus suggested that self-efficacy and mastery
70
orientation can be developed among learners, and learners’ extrinsic motivation towards learning should be suppressed for language learning in light of the negative effect it may bring to self-perceived proficiency. The strengthening of self-efficacy can possibly lead to learners’ confidence in their capabilities to finish tasks.
Self-efficacy should be emphasized in particular, in terms of the strongest predictive power it had on self-perceived proficiency. Additionally, if learners are stimulated to learn for the mastery of the target language (i.e. mastery orientation) rather than for the performance manifested to others (i.e. performance orientation), they may have more confidence in their own proficiency.