It depends on what kinds of statistics you are talking about. If you want to test hypotheses and evaluate correlations, etc. than an N= 30 is almost certainly too small to produce meaningful results. The reason is because your standard errors will be quite large and confidence intervals correspondingly quite wide (in other words, your tests will have very low power).

It depends on what kinds of statistics you are talking about. If you want to test hypotheses and evaluate correlations, etc. than an N= 30 is almost certainly too small to produce meaningful results. The reason is because your standard errors will be quite large and confidence intervals correspondingly quite wide (in other words, your tests will have very low power).

Around 30 may be enough for some analysis according to the central limit theorem, but yes it depends on the kinds of statistics and variables. 

Even though 30 is enough but it might be too small to obtain any significant finding

Dear Khondokar, I agree with my colleagues, 30 person for statistical analysis maybe enough in some cases, but it will be difficult to get significant results and overview the subject in-depth.