Two-way ANOVA help
Hi, so a rundown of what my experiment is: I conducted separate 96 hour duration toxicity tests with three different life stages of a benthic amphipod: neonate , juvenile, and adult. The tests were exposing them to various mercury concentrations to assess mortality. So my two way anova was conducted with two factors: life stage and mercury concentration on mortality. I kept reading that arcsin transform on percent mortality data is used with these types of tests to improve normality and variance so when I arcsin transformed my mortality data I ended up getting that my normality (Shapiro-wilk) failed (p < 0.05) and equal variance (brown-forsythe) also failed (p < 0.05)
Now this is where my question arises …
I decided to run the anova with the raw , untransformed mortality data and it still failed normality but passed equal variance (p = 0.076)
Do I just go with the raw data and then say that even though normality was violated, results were retained due to robustness of ANOVA?
Also, one last thing I did try and log transform (ln) my mortality data and run the ANOVA but it didn’t really work with my data
Any feedback would be appreciated plz and thank you 😭