nolanheinzen3e

Reporting Newer Statistics

It is no longer enough to simply present the descriptive statistics and the results of the hypothesis test. As of the 2010 edition of its Publication Manual, the APA requires that effect sizes and confidence intervals be included when relevant. The effect-size estimate is often included as part of the report of the statistics, just after the p value. There is often a statement that indicates the size of the effect in words, not just in numbers. The confidence interval can be presented after the effect size, abbreviated as “95% CI” with the actual interval in brackets. Note that nonparametric tests often do not have associated measures of effect size or confidence intervals. In these cases, researchers should provide enough descriptive information for readers to interpret the findings.

To summarize this aspect of the Results sections, we include:

Effect size, along with a statement about the size of the effect.
Confidence intervals when possible, along with a statement interpreting the confidence interval in the context of the study.

For the study on humor, we might report the effect size as part of the traditional statistics that we described above: There was not a statistically significant effect of gender, t(7) = −0.03, p = 0.98, d = −0.02, 95% CI [−28.37, 27.67]; this was a small, almost nonexistent, effect. In fact, there is only a 0.35% difference between the mean percentages for women and men. This study does not provide evidence that men and women, on average, rate different percentages of cartoons as funny.

For the humor study, we can now pull the parts together. Here is how the results would be reported:

To examine the hypothesis that women and men, on average, find different percentages of cartoons funny, we conducted an independent-samples t test. The independent variable was gender, with two levels: female and male. The dependent variable was the percentage of cartoons deemed funny. There was not a statistically significant effect of gender, t(7) = −0.03, p = 0.98, d = −0.02, 95% CI [−28.37, 27.67]; this was a small, almost nonexistent, effect.

Based on the hypothesis test and the confidence interval, this study does not provide evidence that women (M = 82.25, SD = 17.02) and men (M = 82.60, SD = 18.13) deem, on average, different percentages of cartoons to be funny. In fact, there is only a very small difference between the mean percentages for women and men, just 0.35%.