Your question is somewhat vague. Are you trying to refer 'data' as 'sample size'?? Do each group consist of 20 data values?? let me try to respond as I understand your question. Your sample size depends on the type of data you are handling. The minimal data or sample size should generally be at least four for each group in order to make statistical comparison. Sample size for each group may be consists of the same or different no. of values. The more sample size you have, the more reliable results you will get. But you have to choose an appropriate statistical tool to better handle larger data.