Entering edit mode
9 months ago
f6f66bd0
•
0
can anyone guide me on how to on how to statistically approach my datasets from a knockout. I’m unsure of what variables I should be comparing / most appropriate statistic test
- Seizure duration; dependent ( continuous) Strains; independent (CS control, CS KO, eas control, eas KO) (categorical)
- Time values independent (continuous)
- Three different datasets in different experimental conditions: light dark (LD), constant light (LL) and constant dark (DD) ( independent, categorical)
Research hypothesis: KO is associated with diurnal seizure variability
Sample size n = 20 biological replicates (strain) per time point
Data doesn’t have normal distribution, tried normalising and still didn’t pass Shapiro test
I’ve used mixed effects despite normalisation and non parametric tests
The distribution should not be normal as duration is non-negative. Try using a generalized linear (mixed) model with a more appropriate distribution like lognormal, gamma or others that might fit your data.