本研究針對183位國小二年級學童進行為期三年的國語文能力成長組型探討，希望能對國小學生國語文能力發展的趨勢及其與認知變項的關聯提供一份初步實徵的全觀視野。文中國語文能力分為聽力與書面兩個分測驗，每年各進行一次施測，每位學童共有三個時間點六項能力估計。研究中依據國語文能力測驗二、三、四年級全省橫斷性研究中所建立常模的能力分配特徵，由原183位樣本擇取100位形成模擬常模，作為線性關係強度和不同成長組型出現比率推估等議題探討的資料基礎。報告中成長組型的分類依據為學童跨三年在兩項國語文能力的剖面，就其可意義化和組型的明確程度，初步的將模擬常模中75%的學童分成五種不同的成長類型，其中高、低國語文能力兩組分別常模的25%和16%。五組中以閱讀低落組學童(佔10%)的成長剖面最為獨特，該組學童聽力和書面兩測驗相對成就水準的落差很大;參照於其他四組，閱讀低落組聽力測驗表現大幅頗於書面測驗，而且他們跨三年在書面測驗的成長斜率異常平緩，即沒有明顯的成長。而在模擬常模中，閱讀低落學童的出現率，男生遠高於女生(7:1)。由模擬常模之國語文能力與各項認知能力的相關組型顯示，認知變項對國語文表現變異的預測力大致在40%到60%之間。文中進一步以認知變項鑑別有待補救的能力和閱讀低落兩組學童，二、三、四年級正確歸類的比，低能力組依序為69%，75%和67%;閱讀低落組可被鑑別的比率則偏低。整體而言，低能力組可在二年級進行鑑定，而本研究的認知變項尚未能有效鑑別閱讀低落組學童。如果選取鑑別功能較大的變項，如視聽覺的短期記憶或工作記憶(包括音韻處理或轉錄)等變項配合畢保德語彙測驗，對低能力組也可達到44%到60%鑑別功能。

　　　　　 The purpose of this study is to investigate the characteristics of language learning progress of elementary school children to enrich the empirical references of screening and intervention decisions of the Chinese learning disability. One hundred and eighty-three students were included for this longitudinal study. One hundred subjects were selected to mirror the characteristics of the original provincial norm. Both written and listening performances were assessed at the end of the second, third and fourth grade. The parameters of the tested items were scaled by IRT (item response theory) concurrent calibration and linear transformation processes. Basing upon six scaled ability estimates, five learning progress patterns were derived. About 25% and 16% of the simulated norm were successfully classified as sharp and slow language learners respectively. Around 10% students were distinctively classified as the weak reading learners (WRL). Compared to the other groups, the reading comprehension achievement levels of WRL was significantly below their listening comprehension. Their growth profile of reading comprehension was relatively flatter than the profiles of the other groups. The ratio of male and female WRL is 7 to 1. The variables of short term memory, working memory, and oral vocabulary correlated with students' language achievement consistently across three years. The cognitive variables can account 40% to 60% amount of variance of language achievement. These cognitive variables can also discriminate 70% slow language learners at second grade. More data will be needed to validate the derived learning progress patterns. Also, the information of affective aspect of the WRL will be very helpful for further intervention designs.