I've been using dual luciferase assays to study promoter activity. Each experiment consists of 6 technical replicates performed in triplicate, with all data normalised to a positive control (tk-renilla plasmid). As the data violates the normality and equality of variance assumption for parametric ANOVA, I've used Welch's ANOVA. I'm getting consistent and significant results, HOWEVER, I have not excluded any outliers (there are some).
Does anyone know of a standard procedure for excluding outliers from dual luciferase data? And if not, given the results are highly significant, across several replicates, is the exclusion of outliers necessary, or even a good thing to do, when the dual luciferase assay is inherently quite variable?
Many thanks in advance for any tips,