0b2328fd-nificant-what-is-the-correct-interpretation-for-reporting-in-the-result

Question

A clinical trial with k = 4 treatment arms yields a significant ANOVA. Tukey-Kramer results: significant differences are found only between Treatment D and each of A, B, and C (three significant comparisons); none of A vs B, A vs C, or B vs C are significant. What is the correct interpretation for reporting in the results section?

Accepted Answer

Report that the omnibus ANOVA indicated differences among groups and post hoc Tukey-Kramer tests localized the difference to Treatment D: D differs significantly from A, B, and C, while A, B and C do not differ from each other.

Answer

Report that the omnibus result is not interpretable because more than one pair is significant; recommend additional Bonferroni-corrected t-tests to confirm all pairwise differences.

Answer

Report that the ANOVA was misleading and only the pairwise results matter; therefore, ignore the ANOVA and conclude all treatments are equivalent except D which must be excluded from all further analyses.

Answer

Report that because multiple pairwise tests produced significant results, it is likely that A, B and C are each significantly different from one another even if Tukey-Kramer did not detect those differences.