I think this question needs more elaborations, but yes it is possible to use the t-test if you have only two groups being compared and if the two groups are independent of each other. If you are, however, talking about the same group under different scenarios, you need to use a paired t-test instead. Also, if the groups are more than two and the data are close to normal or bell shape distribution with uniform variances, you need NOVA followed by a means separation (preferably Tukey's HSD, Duncans MRT, or Dunnets if there is a check (control) involved. I hope this helps.