I am wondering if it is possible to skip missing value imputation. I am forced to complete this step even if there are not differences in my groups. I have the feeling that the imputation is affecting to my results and I would like to be able to perform the statistical analysis considering the N/A (and not an artificial number). Is that possible in any way?
Missing value imputation is not related to whether there is difference or not in your groups. The main reason is that missing values cannot be handled in most downstream analyses.
Imputation affect your results? This should not be the case - the only reason I can imagine is that your data contains a large portion of missing values …. then if you choose LoD, mean, or PCA-based method will likely have a large, yet different impact on your overall data structure. If this is indeed the case, you must deal with missing values carefully, rather than ignore this
This topic was automatically closed after 3 days. New replies are no longer allowed.