This is a tough one. I suggest you use Lee Becker's pooled variance program (use the hyperlink from a few slides back). Let's say you have a study looking at the DV "Time to help". Your IV is whether potential helpers did not see (condition #1) or saw (condition #2) someone else helping. The mean time to help for those in condition #1 was 58 seconds (variance was 2.81). The mean time to help for those in condition #2 was 55 seconds (variance was 3.06). First, tell me what Cohen's d (pooled) is. Second, tell me if the effect size is meaningful in terms or magnitude!