Statistical methods appear appropriate; however, no evidence homogeneity of variance tests conducted prior to ANOVA or other tests. The data from days 1 to 4 of the test were subjected to general linear model with a split-plot design and pairwise testing using Duncan’s test. Water maze data was repeated over time so a repeated measures MANOVA was more appropriate. Comparison between the performance of the last trial on day 4 vs. the first trial on day 5 was submitted to a paired t test. No other threats to internal validity were identified.

Viberg et al. 2003

Risk of bias

Animal bioassay experiments

HAWC