Some Effects of Including Effect Size and “what if” Information

The authors investigated whether 2 editorial policies concerning the reporting of statistical information would affect readers' perceptions of research results. In 4 experiments, undergraduate and graduate students read fictional research abstracts or articles that included effect size information, “what if” information, or no additional information. Regardless of whether the size of the effect was large or small, undergraduates perceived research results to be more important when effect size information was included and were more likely to believe (a) that a statistically nonsignificant effect would have been significant if only the sample size had been increased, and (b) that the sign of the nonsignificant effect was confirmable when “what if” information was included. Graduate students were less affected by effect size and “what if” information. The authors discuss recommendations concerning editorial policies on statistical information.