Wagner E E, Alexander R A, Roos G, Prospero M K
J Pers Assess. 1985 Dec;49(6):579-81. doi: 10.1207/s15327752jpa4906_3.
Hand Test data from comparable sets of clinical samples were analyzed for reliability using the usual odd-even split and then also calculating all possible (126) split-half combinations. Maximum stable reliabilities were obtained which were substantially superior to the odd-even method. It was concluded that projective tests are probably more internally reliable than has heretofore been reported.
使用常规的奇偶分割法，并计算所有可能的（126种）分半组合，对来自可比临床样本集的手部测试数据进行可靠性分析。获得了最大稳定信度，其显著优于奇偶法。得出的结论是，投射测验的内部信度可能比以往报道的更高。
