The problem of discordant results

A researcher is interested in the effect of X on Y so she collects all the available literature on the topic. She organizes all the relevant research into three piles according to their results. On one side she puts those studies reporting results that were statistically significant and positive. On the other side she puts those studies reporting results that were statistically significant and negative. In the middle she puts those studies that reported results that were statistically nonsignificant. She is unable to draw any conclusions from these disparate results and decides that this is a topic in need of a first-rate study to settle the issue. She conducts her own study and observes that X has a significant negative effect on Y. She writes that her result is consistent with other studies that observed the same effect. However, she is not sure what to make of those studies which found something completely different so she makes some vague comments about “the need for further research before firm conclusions can be drawn.” In the back of her mind she is a little disappointed that she was unable to settle the matter, but she has little time to reflect on this as she is already planning her next study.