2013-08-012024-05-17https://scholars.lib.ntu.edu.tw/handle/123456789/683642摘要：在外語測驗領域當中，已有極多的研究探討了焦慮與測驗表現的相關性。這些研究對焦慮的本質主要採取三種不同觀點：特質觀點、狀態觀點及情境觀點。然而，各觀點所獲取之研究結果皆相當不一致。再者，截至目前為止，尚未有學者整合此些研究以獲取更為可信之結論。是故，本研究將建立一理論模型來表達四種不同焦慮〔特質焦慮、狀態焦慮及二種情境焦慮（外語焦慮及測驗焦慮）〕與外語測驗表現之相關性，並透過後設徑路分析來檢驗此模型之適切度。主要目的有二：一為獲取並比較該四種焦慮對外語測驗表現之影響的合併效果量，藉此釐清焦慮於外語測驗中之角色；二為深入剖析這四種焦慮間之相互關係。 在此研究中，後設徑路分析進行的主要步驟有二。第一，研究者將透過後設分析來為所有變數建立一相關係數矩陣。亦即對每一組變數，研究者將：(1)界定其操作定義，(2)搜尋主要研究，(3)評估各主要研究的可納入性及為其編碼，(4)計算合併效果量以作為該組變數之相關係數及偵測出版偏誤存在之可能性。第二，研究者將對該相關係數矩陣進行徑路分析，藉此檢驗本研究所建構之理論模型的適配度以及估算相關徑路係數。 最後，研究者將根據研究結果來對外語測驗理論、研究方法及教學實務提出建議。 <br> Abstract: In the field of second/foreign language (L2) learning and testing, a remarkably extensive body of research has hitherto explored the role of anxiety in contributing to the learning and performance variations on the part of the L2 learners and test-takers. Moreover, this bulk of research has been considered to take three different perspectives on the nature of anxiety, viz., trait, state, and situation-specific perspectives. However, a close examination of the studies in each perspective that investigated the relationship of anxiety with L2 test performance brought to light inconsistent patterns for this relationship. Furthermore, thus far, no research effort has been made to synthesize the quantitatively-oriented L2 studies exploring this anxiety-performance relationship. Drawing on the inconsistent patterns of the relationship between anxiety and L2 test performance and the absence of relevant research synthesis, the current research project thus aims to employ the meta-analytic path analysis procedure to test a theoretical causal model proposed to characterize the interrelationships among four different types of anxieties (viz., trait anxiety, state anxiety, language anxiety, and test anxiety) and L2 test performance. In so doing, it attempts to bridge the gap in the extant research literature by revealing the synthesized effects of anxiety on L2 test performance, comparing the relative strength of the effect of the four targeted anxieties on L2 test performance, and shedding light on the meta-analyzed interrelationships among these anxieties. Capitalizing on the meta-analytic path analysis technique, the current researcher plans to test the theoretical causal model proposed in this academic project by taking two major steps. In the first step, he will adopt the meta-analytic procedure to generate a correlation matrix for the five targeted variables, namely, trait anxiety, state anxiety, test anxiety, language anxiety, and L2 test performance. Specifically, in this step, for each pairwise relationship, he will (1) operationalize the two variables, (2) perform a comprehensive search for the relevant primary studies, (3) evaluate the obtained primary studies for inclusion based on the eligibility criteria and code the relevant features of each of the remaining primary studies in accordance with a pre-established coding scheme, and (4) calculate individual effect sizes, synthesize these effect sizes following the random-effects model to attain an aggregate effect size to serve as the meta-analyzed correlation estimate of the two variables, and check for the evidence of publication bias. In the second step, he will utilize the path-analytic procedure to test the proposed model by using the correlation matrix generated in the first step. Specifically, he will path-analytically evaluate the goodness-of-fit of the proposed model with reference to multiple fit indices and estimate the relevant parameters of the model. In light of the resultant findings, the researcher will put forth implications for L2 testing theory, research methodology, and instructional practices.後設徑路分析焦慮外語測驗表現Meta-analytic Path AnalysisAnxietyL2 Test Performance