中华成人智力量表T分转换系统的建立和效度检验

Development and validation of T-score conversions for the Intelligence Scale for Chinese Adults

目的：为中华成人智力量表（ISCA）建立T分转换系统,将测验粗分转换为人口统计变量校正的T分,并检验其效度.方法：基于ISCA常模资料为ISCA各分测验分数建立以年龄、性别、受教育程度和职业等人口统计变量为自变量的线性回归模型,将残差转换为均数是50、标准差是10、取整的分测验T分.累加6种不同组合的分测验T分,并线性转换为均数是50、标准差是10、取整的组合T分.计算95名正常成人（正常组）、102名脑梗死患者（脑梗死组）及与其人口统计变量匹配的健康配对者（配对组）的T分,检验T分的效度.结果：生成T分的回归模型均有统计学意义（均P ＜0.001）.正常组的分测验T分平均为46.8 ～51.3分,没有残留人口统计变量主效应.不同受教育程度者之间的分测验T分异常的人数比例差异无统计学意义.脑梗死组分测验T分平均数为40.7 ～47.8分,配对组为48.2 ～51.7分.脑梗死组有10项分测验T分和6项组合T分均低于配对组（均P＜0.01）,脑梗死组7项分测验和6项组合T分40分的人数比例高于配对组（均P＜0.01）.结论：本研究为ISCA建立了T分转换系统.生成的T分与人口统计变量没有显著关联,能有效区分认知损害与正常,提示T分转换系统具有一定效度.

Objective：To develop T-score conversion system for the Intelligence Scale for Chinese Adults （ISCA） and test its validity,in order to help clinicians convert raw scores on subtests of the ISCA into T-scores correcting for age,gender,education and occupation.Methods：Linear regression analyses were conducted with the standardization data of the ISCA and models were established to estimate scores on subtests of the scale with demographical variables.Residuals of regression models were transformed into integral T-scores on the subtests with mean of 50 and standard deviation of 10.Six distinct composite T scores on the full scale and five subscales were derived from sums of T-scores on subtests submitted to them,and they were also with the mean of 50 and standard deviation of 10.T-scores for 95 healthy adults （healthy group）,102 patients with cerebral infarction （patient group）and 102 healthy adults demographically-matched with the patients （match group） were computed with the conversions.Results：Regression models for computing T-scores for all subtests were all statistically significant （Ps ＜0.001）.The mean T-scores on the 12 subtests of the healthy group were 46.8-51.3.No significant demographical effects were found in these subtests T-scores.There were not significant differences among three educational levels regarding the rates of persons with abnormal T-scores.The mean T-scores for the patients group were 40.7-47.8 while those for the match group were 48.2-51.7.Mean T-scores on 10 subtests and all 6 composite T-scores in the patients group were significantly lower than those in the match group （Ps ＜0.01）.Accordingly,percents of the patients with abnormal T-scores were larger than those in the match adults on 7 subtest T-scores and all composite T-scores （Ps ＜ 0.01）.Conclusion：In this study,the T-score conversion system for the ISCA has been developed.There are no significant association between T-scores and demographical variables.Cognitive impairment could be identified effectively based on the T-scores.These findings suggest that the T-score conversions are valid.