Anavex Life Sciences Corp (AVXL)

[Doc328]

Our random-selection trial had 33% of patients on placebo, 66% dosed. In the left-side chart, this is actually the same distribution for all "top interval" episodic memory improvers at week 14. In other words, no apparent effect between the two cohorts (as portrayed in this manner).

Yes n values in the graph or legend would have cleared this up.

And for this slide we use the J-T test instead of ANOVAs, as there were 3 samples to compare. With ANOVA and doing 2 sample testing, our p-val may have been over .05. You can see how this is likely when you note the p-value for slide 18, where we do use ANOVA

Another good point, data like this is often parametric so ANOVA might have given a different p than a nonparametric test. Likewise, using the Kruskal-Wallis is far more standard than the J-T (I had to look this one up and won't try to spell it). The K-W would have given a higher p.