I am doing stats analysis on the quantification of DNA damage under two conditions (light and dark) across 4 time point, the quantity was reduced across time (which seems normal) and some may have decreased to 0 at the third time point, which also 0 in the fourth time point.
Then i used MANOVA analysis for my study, amount of damage under light and dark as my DVs and time as my IVs.
When i did my MANOVA test, the box's test of equality of covariance matrices was not computed because there are fewer than two nonsingular cell covariance matrices.
Then i checked the levene's test of homogeneity of covariance, it is violated in one of the two dependent variables.
Then i checkd the normality test and it is also violated.
May i know which analysis i should use?
Hoi,
Kruskal-Wallis is nonparametric equivalent to M/ANOVA, but if your baseline (T0) values are normally distributed then you can group-mean center your variables and conduct time-series regression for level-1 only change.
Hope this helps,
Matthew
If you are familiar with R, maybe the nparLD package can help. I think you have a f1.LD.f1 design (1 between x 1 within). You might also want to try the npIntFactRep package, which is for 2-way mixed design (it only outputs the p-values for the interaction, not for the main effects). The data input is in 'wide' format for npIntFactRep but in wide format for nparLD. See tthe example in attachment.
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication