6.3 Results of Statistical Tests

Here are some guidelines for reporting the results of statistical tests:

  • Your Methods section should clearly state the significance level (\(\alpha\)), and this should be decided prior to the study

  • Test statistics (e.g. Student’s t or an F from ANOVA) should be rounded to 2 decimal places, and associated P-values should report 3 decimal places, or if smaller than 0.001, then <0.001. P-values do not indicate effect size, so reporting \(P = 10^{-6}\) is not more impressive than P <0.001

  • Concluding statements should, in the absence of a table, include the test, test statistic value, degrees of freedom df or sample sizes, P-value, and confidence interval (if appropriate) in parentheses.

  • For example: "On average, hair loss was significantly greater among fathers compared to childless men (Welch’s 2-sample t-test; t = 4.23; n F = 18, n C = 20; P = 0.018; 95% CI for difference: 9.34 – 18.22%)."

  • Regression and ANOVA results should be shown in a standard ANOVA table format.

Example of ANOVA table format

Table 3.

SS df MS F P
Treatment 7.224 2 3.6122 7.29 0.004
Error 9.415 19 0.4955
Total 16.640 21 4.1078