目的︰閱讀障礙的成因咸認為是多重因素造成的，本研究之目的在探討閱讀障礙學童與正常學童之間其氣質組合亞型是否有差異。方法︰首先以標準化之閱讀成就測驗對國小二年級學生施測，選取智商高於85者，並排除嚴重身心疾患且成就測驗得分低於常模之百分10位值的學童67名為閱讀障礙組，另選取智商高於85者，並排除嚴重身心疾患且成就測驗得分介於常模平均值上下一個標準差之學童67名為對照組。兩組學童分別以「兒童氣質量表」評估其氣質特徵，並按Carey及McDevitt氏分類法，依規律性、趨避性、適應度、反應強度、情緒本質等各分項得分之組合，將每一學生歸類於五個氣質組合亞型之一。結果︰閱讀障礙組59名，對照組64名，共123名完成評估，以卡方檢定檢驗兩組在五個氣質組合亞型的分布差異，發現兩組間有顯著的差異（P=0.0134），其中慢吞吞型氣質組合亞型僅見於閱讀障礙學童。結論︰慢吞吞型氣質組合亞型學童比起其他氣質組合亞型學童更易有閱讀障礙之問題。著者以goodness of fit的概念來討論氣質與學習成就間可能的關係，即針對不同氣質亞型閱讀障礙兒童應考慮施予不同之矯治教育。

　　　　　Objectives: Reading disability is thought to have a complex multifactorial etiology which is poorly understood. Further investigation of the etiology and presentation of reading disability may allow improvements in the design of special education programs. This study examined whether temperamental differences exist between reading disabled and normal 2nd grade elementary school students in Taiwan. Methods: Students were first tested individually with a standardized reading achievement test. The experimental group consisted of sixth-seven of these students who had reading achievement scores below the 10th percentile of the normal distribution curve, IQ above 85 and no obvious neurological or mental abnormalities. The control group consisted of 67 students who had reading achievement scores within the range of one standard deviation from the mean. Students in both groups were assessed by the Chinese Children's Temperament Questionnaire (CCTQ) and temperamental subgroups were classified into temperamental subgroups according to scores for rhytmicity, approach-withdrawal, adaptability, quality of mood, and intensity of reaction. Results: The results showed that there were differences in the distribution of temperamental subgroup classifications between reading disabled students and normal controls (p=0.00134). Most importantly, the slow-to-warm-up subtype was only found among the reading disabled group. Conclusion: Children with slow-to-warm-up temperament were more apt to be reading disabled than children with other temperamental subtype classifications. The goodness of fit between reading disability data and temperament classifications is discussed.