A researcher performs a one-way ANOVA with α = 0.05 and obtains p = 0.12 (fail to reject). She nevertheless conducts Bonferroni-adjusted pairwise tests and reports one significant pair. According to standard practice, what is the correct interpretation?

• A. Because Bonferroni adjusts for familywise error, the significant pairwise result completely overrides a non-significant ANOVA and indicates that factor explains variance even if F was not significant.
• B. She should not place strong confidence in the pairwise significance because the overall ANOVA did not show evidence that any mean differs; the post hoc result may be a chance finding.
• C. She must rerun the ANOVA with a different alpha level until the omnibus test becomes significant, and then accept the pairwise significance once the F-test is significant at that new α.
• D. The ANOVA result is irrelevant; any significant pairwise test after Bonferroni adjustment is valid and proves there is a true pairwise difference even though the omnibus test failed.

Answer: (B)