Abstract: Reading comprehension is the ultimate aim of learning to read, and it depends primarily on word reading during early school years. On the basis of the moderate or strong correlation between word reading and reading comprehension, more and more recent studies argued that the growth rate of word reading also contributes to reading comprehension, and it was confirmed by some studies on reading acquisition in alphabetic languages. However, there are few studies examining the contribution of the growth rate of word reading to reading comprehension in Chinese acquisition. Moreover, few studies compared the relative contribution of two word reading skills, i.e., word decoding and reading fluency, to reading comprehension. Therefore, the aim of this study was twofold. The first was to examine whether the initial status and growth rate of Chinese word reading (word decoding and reading fluency) during Grade 1 uniquely contributed to reading comprehension at the end of Grades 1 and 2, and the second was to compare the relative contributions of word decoding and reading fluency to reading comprehension. One hundred and eighty-three Grade 1 students (101 boys and 82 girls) participated in this study, and they were assessed at the November, February, and May during Grade 1 (T1, T2, and T3), and at the May of Grade 2 (T4). They were assessed on nonverbal IQ (T1, Nonverbal Matric), word decoding (T1-T3, Character Recognition), reading fluency (T1-T3, One-Minute Reading), and reading comprehension (T3-T4, Passage Comprehension). Hierarchical linear modelling was used to estimate the intercepts and slopes of word decoding and reading fluency based on the data at three time points (T1-T3). Results showed that the intercept correlated with the slope for both word decoding (r = -.65) and reading fluency (r =-.52). Results of hierarchical regression analysis showed that the slope of word decoding uniquely predicted reading comprehension at T3 and T4, after nonverbal IQ and the intercepts of word decoding and reading fluency were statistically controlled. Moreover, the slope of reading fluency uniquely predicted only reading comprehension at T4, but it made unique contribution to the growth of reading comprehension from T3 to T4 (by controlling the autoregressive effect). The contribution of word decoding (the intercept and slope) to reading comprehension at T3 was higher than that of reading fluency (the intercept and slope), while their contribution to reading comprehension at T4 was approximately equivalent. The results indicated that the growth rate of word reading is important for the development of reading comprehension in Chinese acquisition, and the policy makers and reading educators should take the growth rate of Chinese word reading into consideration when using word reading to predict later reading achievement. The increasing contribution of reading fluency to reading comprehension also confirmed the theory of automatic information processing in reading, which argued that the automaticity of word decoding reallocated the attention resource from basic word recognition to higher-order comprehension processes.

Key words: Reading comprehension, Word decoding, Reading fluency, Hierarchical lineal modelling, Initial status, Growth rates

摘要： 本研究测查了183名儿童在一年级期间（学年初，学年中，学年末）的词语阅读能力（词语解码，阅读流畅性）及一年级末和二年级末的阅读理解能力，采用分层线性模型估计每名儿童在一年级期间词语阅读能力的初始水平和发展速率，以考察二者对阅读理解能力的独立预测作用。结果表明，词语识别的初始水平和增长速率都能独立预测一年级末和二年级末的阅读理解能力，阅读流畅性的初始水平和增长速率只能独立预测二年级末的阅读理解能力，但可以预测阅读理解能力在一年级到二年级的增长。

关键词: 阅读理解能力, 词语解码, 阅读流畅性, 分层线性模型, 初始水平, 发展速率