摘要:
本研究采用多元概化理论中的p × i随机测量模式的设计，对503名大学生进行测量，以考察成就动机量表的测量信度，从而为修订该量表提供依据。结果表明：多元概化G研究结果表明，成就动机量表的两个分量表相互独立。D研究结果表明，两个分量表的概化系数均在0.70以上，达到了可接受水平。在测验信度要求较低的情况下，当追求成功、避免失败分量表题目数量分别为10、7时，成就动机量表的测量信度达到可接受水平。
Abstract:
In this study, a total of 503 college students were surveyed with the Achievements Motive Scale (AMS), with the p × i random pattern design of multivariate generalizability theory (MGT) to explore the psychometric properties of the Achievements Motive Scale, in order to provide the basis for re-vising it. The results as follows: G study shows that the two subscales were independent. D study shows that both of the two subscales’ generalizability coefficients were more than 0.70. When the Ms and Mf has 10 and 7 items respectively, the Achievements Motive Scale’s reliability could be ac-cepted in the condition of low reliability.