本研究探討華語二語學習者之詞彙知識與構詞覺知於閱讀理解之影 響，研究對象為 34 位中級以上之華語二語學習者。初步研究發現：（1） 學習者之詞彙知識與各文本閱讀理解皆顯著正相關，如閱讀理解表現分別 和詞彙廣度知識與深度知識有顯著高度正相關（r ＝ .871，p ＜ .001；r ＝ .790， p ＜ .001），廣度知識對閱讀理解有較大解釋力（R 2 ＝ 0.759）；加入學習華 語期間為控制變項時，學習華語期間能解釋閱讀理解中 17.8%變異，廣度 知識能解釋另外 58%變異，閱讀理解和構詞覺知則無顯著相關；（2）廣度 知識分別與構詞覺知和複合詞覺知有顯著正相關（r ＝ .353，p ＝ .041；r ＝ .353 ，p ＝ .041），與派生詞覺知則無；構詞覺知可解釋廣度知識 12.5%變異， 複合詞覺知則可解釋 12.4%；（3）詞彙廣度知識與詞彙深度知識有顯著高 度正相關（r ＝ .856，p ＜ .001），廣度知識可解釋深度知識 73.2%變異。
This study aims to examine the impacts of vocabulary knowledge and morphological awareness on reading comprehension of L2 Chinese learners. Thirty-four CSL learners of intermediate level and above participated in this study. The current findings revealed the following results. First, CSL learners' vocabulary knowledge was highly correlated to overall or respective text reading comprehension. Furthermore, reading comprehension was highly related to breadth and depth of vocabulary knowledge of CSL learners, respectively (r = .871, p < .001; r = .790, p < .001), while the breadth of vocabulary knowledge was a more powerful indicator for reading comprehension (R2 = 0.759). When adding the factor of the learning time to the model, 17.8% of variance of reading comprehension could be explained by the learning time, while another 58% of variance could be further explained by the breadth of vocabulary knowledge. Nevertheless, the correlation between reading comprehension and morphological awareness was not significant. Second, learners' breadth of vocabulary knowledge was moderately and significantly related to morphological awareness and compound awareness respectively (r = .353, p= .041; r = .353, p = .041), while there was non-significant correlation between breadth of vocabulary knowledge and derivational awareness. Regarding breadth of knowledge, 12.5% of variance could be explained by morphological awareness, and 12.4% of variance could be explained by compound awareness. Third, L2 learners' breadth of vocabulary knowledge was significantly related to the depth of vocabulary knowledge (r = .856, p < .001), and 73.2% of variance in depth of knowledge could be explained by breadth of vocabulary knowledge.