If cases of different diagnostic criteria or of different duration and from different settings were chosen, they are incomparable as differences exist in-between them.

It is true with controls also.

If the cases were selected from hospitals and compared with the controls from general population, the accuracy of the study will suffer.

Selection of cases and controls from different sexes or occupations will also deviate the results from truth.