5.2.1 Language Achievement & Metalinguistic Knowledge
Based on the correlation analysis, metalinguistic knowledge had a very strong correlation with language achievement (r = 0.6). As for the hierarchical regression analysis, it was found that through Model 1 to Model 5 of Table 5 metalinguistic knowledge had a significant predictive effect on language achievement when the effect of self-perceived proficiency was controlled for and other independent variables (i.e. motivational components, self-regulation and self-perceived proficiency on the four language skills) were included in the analysis. Additionally, metalinguistic knowledge could also predict language achievement significantly even when the effect of self-perceived proficiency was not controlled for (β = 0.6, p < .01). It suggests that metalinguistic knowledge is a strong predictor of language achievement even when the motivational components, self-regulation and self-perceived proficiency on the four language skills are incorporated altogether into the analysis.
The results are in accordance with the positive relationship between
71
metalinguistic knowledge and L2 performance found in previous metalinguistic studies (e.g., Renou, 2001; Roehr, 2007; Morris, 2003). The results revealed in the present study not only suggest that learners’ metalinguistic knowledge is highly related to their language achievement but also that learners’ metalinguistic knowledge can predict their language achievement significantly. Indeed, learners’ performance on monthly exams largely depends on their metalinguistic knowledge since monthly exams tend to focus on evaluating their knowledge and mastery of the subject matter.
It requires learners’ application of sufficient metalinguistic knowledge to answer the test questions. Learners have to be equipped with knowledge about grammatical terms, rules, and relations among these grammatical categories (i.e. their metalinguistic knowledge) if they anticipate a good performance on the monthly exam (i.e. language achievement). Metalinguistic knowledge seems to play an essential role in language achievement considering its much higher predictive power than the other variables.
The findings can provide pedagogical implications for teachers that in order to help learners obtain satisfactory language achievement, teachers can enhance learners’
metalinguistic knowledge by highlighting the importance of grammatical terms and rules in a meaningful context in addition to assisting learners with their understandings of the text meanings. Learners’ explicit knowledge about the language (i.e. metalinguistic knowledge) can be incorporated into formal instruction with a view to helping learners become more familiar with the technical terminology of the target language and gradually the mastery of the language can be achieved.
5.2.2 Language Achievement & Self-perceived Proficiency
The correlation analysis indicates that language achievement correlated with self-perceived proficiency significantly (r = 0.42) as well as with self-perceived proficiency on the four language skills. The significant relationships between
72
language achievement, self-perceived proficiency and self-perceived proficiency on the four language skills make it necessary to carry out the hierarchical regression analysis by entering the independent variables in sequence to examine the true effect that the other independent variables may have on each of the dependent variables.
Correlation analysis alone cannot be counted as the valid representation of the effects of the independent variables on language achievement and self-perceived proficiency since pseudo correlations may be presented and the dependent variables may confound each other (as judged from the significant relationship between them) if they are included in correlation analysis altogether.
According to Model 5 of Table 5, when self-perceived proficiency was divided more specifically into the four language skills, only self-perceived reading ability held a significant predictive effect on language achievement (β = 0.15, p < .05) as metalinguistic knowledge, motivational components and self-regulation were included in the analysis.
The results suggest that participants’ self-evaluation of their reading proficiency is highly related to their language achievement. Learners who have confidence and belief in their own reading ability tend to have good performance on exams since many exams aim to evaluate learners’ reading abilities through vocabulary test, reading comprehension questions or cloze items and learners have to read and understand the test items before they provide answers. Learners’ reading ability plays a critical and decisive role in language achievement. Teachers who expect learners to have better performance on exams can begin by facilitating learners’ reading ability and encouraging them to read more extensively. By doing so, learners may build confidence in their reading ability gradually and that can help them make progress on exams and lead to better language achievement.
73
5.2.3 Language Achievement & Self-regulation
There was a significant correlation between language achievement and self-regulation (r = 0.4) according to the correlation analysis. When self-regulation and metalinguistic knowledge were incorporated into the hierarchical regression analysis with language achievement as the dependent variable in Model 3 and the effect of self-perceived proficiency was controlled for, self-regulation exerted a significant predict power on language achievement (β = 0.16, p < .01). Even when the effect of self-perceived proficiency was not controlled for, self-regulation still held a significant predictive effect on language achievement (β = 0.4, p < .01). In addition, when the five components of motivation were added to the analysis in Model 4, the significant effect of self-regulation on language achievement increased (β = 0.18, p
< .01) compared with its effect in Model 3. The predictive power of self-regulation became stronger in Model 5 when self-perceived proficiency on the four language skills was added to the analysis.
The findings suggest that self-regulation has an important and influential role in language achievement. According to Tseng’s (2008, p.7) study, self-regulation is “the self-directed processes by which systematic operations of thoughts, feelings, and behaviors are attuned to take control over a chosen action.” Self-regulation plays an essential role in language learning because it helps learners to develop responsibility for their learning outcomes (Zimmerman, 1990). Learners with high self-regulation are more responsible for their learning and they adjust and direct their learning process systematically to control learning and achieve their goals. Since self-regulation held a significant predictive power on language achievement, it is probable for self-regulated learners to possess better language achievement because they are able to take control over their own learning until the task is accomplished.
The findings can also account for the phenomenon that learners who take
74
responsibility for tasks and who always regulate themselves against temptations (e.g., TV, games, food, sleep, etc.) until they achieve what they are required do are mostly successful language achievers. Consequently, teachers can arouse students’ awareness of self-regulation to language learning and facilitate the development of self-regulation to help them obtain better language achievement.
5.2.4 Language Achievement & Motivation
The correlation analysis shows that language achievement correlated with motivation significantly as well as with the five components of motivation. As for the hierarchical regression analysis, none of the motivational components could predict language achievement significantly when controlling for the effect of self-perceived proficiency and including metalinguistic knowledge in Model 2. However, self-efficacy (β = 0.412, p < .01) and intrinsic motivation (β = -0.144, p < .05) had significant effects on language achievement when the effect of self-perceived proficiency was not controlled for. When Model 2 and Model 4 are juxtaposed for comparison, Model 4 indicates that the effect of intrinsic motivation increased and became significant (β = -0.14, p < .01) with the entry of self-regulation. The predictive power of intrinsic motivation remained significant in Model 5 when self-perceived proficiency was divided into the four language skills.
As seen from the obvious change in the predictive effect of intrinsic motivation on language achievement, it is possible that self-regulation is a suppressor variable, whose effect is to suppress or alter the apparent relationship between language achievement and intrinsic motivation (Warner, 2008). Self-regulation alters the original predictive effect of intrinsic motivation on language achievement and makes its predictive power significant. Besides, one possible interpretation for the significantly negative effect of intrinsic motivation on language achievement is that
75
intrinsically-motivated learners who learn a language for fun may not have good performance on exams. This may provide a possible explanation to the phenomenon that many students like English very much and learn English out of pure interest but they do not obtain high scores on exams. Intrinsic motivation alone may not be of great help in facilitating learners’ language achievement; on the contrary, intrinsic motivation actually had an adverse effect on language achievement. As mentioned in previous sections, good language achievement may not only be influenced by n intrinsic motivation, metalinguistic knowledge, self-regulation and self-perceived reading ability all contribute to language achievement significantly. It is possible that learners who are equipped with the above characteristics are able to perform well on exams and have satisfactory language achievement. Particularly, metalinguistic knowledge should be emphasized in classroom on account of the fact that it had the strongest predictive power on language achievement.