Crockett D, Clark C, Labreche T, Lacoste D, Klonoff H
J Clin Neuropsychol. 1982 Jul;4(2):167-71. doi: 10.1080/01688638208401126.
Statistical means of shortening the Speech Sounds Perception Test were examined. Stepwise multiple regression analyses indicated that a 30-item version correlated quite highly with the 60-item version. Group means for representative clinical groups were presented, but no strong differences emerged using either version. Inclusion of length of stay and number of previous admissions as covariates did little to enhance the predictive validity of either form. Finally, the correlation of the 30-item test with demographic variables showed that age correlated most, followed by occupational status. these results were interpreted as indicating that the 30-item version could be used while maintaining the clinical utility of this test.
对缩短语音感知测试的统计方法进行了研究。逐步多元回归分析表明，30项版本与60项版本的相关性相当高。列出了代表性临床组的组均值，但使用任一版本均未出现明显差异。将住院时间和既往住院次数作为协变量纳入，对提高任一形式的预测效度作用不大。最后，30项测试与人口统计学变量的相关性表明，年龄的相关性最高，其次是职业状况。这些结果被解释为表明30项版本在保持该测试临床效用的同时可以使用。
