Abstract
Use of pooled interaction error terms in repeated measurement analyses of variance may lead to seriously erroneous conclusions. The danger is increased when designs contain both data sampled at short intervals of time, such as second‐by‐second heart rate (HR) responses, and data sampled at longer intervals, such as responses on different trials or stimulus presentations. Empirical examples illustrate the differing error variances typically found with such data.

This publication has 13 references indexed in Scilit: