Pollatsek A, Well A D
Department of Psychology, University of Massachusetts, Amherst 01003, USA.
J Exp Psychol Learn Mem Cogn. 1995 May;21(3):785-94. doi: 10.1037//0278-7393.21.3.785.
Counterbalanced designs are ubiquitous in cognitive psychology. Researchers, however, rarely perform optimal analyses of these designs and, as a result, reduce the power of their experiments. In the context of a simple priming experiment, several idealized data sets are used to illustrate the possible costs of ignoring counterbalancing, and recommendations are made for more appropriate analyses. These recommendations apply to assessment of both reliability of effects over subjects and reliability of effects over stimulus items.
平衡设计在认知心理学中无处不在。然而，研究人员很少对这些设计进行最优分析，结果降低了实验的效力。在一个简单的启动实验背景下，使用了几个理想化的数据集来说明忽略平衡可能带来的代价，并提出了更合适分析方法的建议。这些建议适用于评估效应在被试间的可靠性以及效应在刺激项目间的可靠性。
