It is much harder to explain nonsignificant findings than those that support our hypothesis. There are two reasons the results of hypothesis tests might not be statistically significant:

1. The effect was not present in the population.
2. The study could not find an effect that is there.

Students often address nonsignificant findings by describing college students as different than the rest of the adult population -- they are smarter, more achievement oriented, more motivated, and more middle class. These characteristics are then used to explain why a particular effect might not be seen in this population. This explanation is often given even though the majority of the research used to develop the hypothesis is conducted on college students.

To explain why effects in the population might not be found in this study, students often discuss the small size of their sample and lack of statistical power. More often than not, this discussion occurs without estimating the test's effect size. Nonsignificant findings in the presence of moderate or large effect sizes indicate a problem of low statistical power. Nonsignificant findings with small effects suggest that the effect may not be present in the population or cannot be detected because of problems with measurement or study design.