I have a dataset that look at the effect on photosynthesis and respiration. When i checked to see if the data looks normal or not, i find that for both the variables the distributio is non-normal. In case of photosynthesis it is obvious because of the effect of difference light. But however for respiration this is not the case and may be because of some outliers. My question now is how to deal with this non-normality of data for both the variables before doing statistical analyses (regression, anova, mixed effect model etc)repiration


  • 1
    $\begingroup$ The bigger issue in the second plot is likely to be unequal variance. But transformation of the other case is not the only solution to this issue -- there are procedures that don't assume normality (GLMs, other parametric approaches, nonparametric approaches). In any case the non-normality is mild and the sample size looks moderately large so it may hardly be an issue at all. But if your models have other predctors than the light/dark then these plots don't even look at the right thing. $\endgroup$
    – Glen_b
    Mar 20, 2015 at 23:49
  • $\begingroup$ The issues surrounding this question have extensive discussions that can be found by searching our site. I have provided a link to a narrow search; remove various keywords to broaden it. $\endgroup$
    – whuber
    Mar 20, 2015 at 23:51
  • $\begingroup$ so i can go ahead without doing any transformation for both the variables? $\endgroup$
    – upendra
    Mar 20, 2015 at 23:51
  • $\begingroup$ I can't tell that from what's here. Perhaps. $\endgroup$
    – Glen_b
    Mar 21, 2015 at 7:23


Browse other questions tagged or ask your own question.