Tukey HSD Post Hoc Test

Tukey's HSD (honestly significant difference) test is a post hoc procedure that compares all pairs of group means after ANOVA while controlling the familywise error rate.

튜키 HSD 검정은 분산분석 이후 모든 집단 평균 쌍을 비교하면서 집단전체 오류율(familywise error rate)을 통제하는 사후검정 방법이다.

After a significant ANOVA, how do I find which groups differ from each other?

분산분석이 유의할 때 어느 집단끼리 다른지 어떻게 알 수 있나요?

Tukey's HSD test compares every pair of group means in a one-way design and keeps the familywise error rate at the chosen alpha level; it is often run after a significant ANOVA, but it does not require one. It fits when group variances are similar; with unequal variances the Games-Howell test is preferred.

When to use Tukey HSD Post Hoc Test

• A one-way design with three or more groups (a significant omnibus F test is customary but not required)
• All pairwise comparisons of group means are of interest
• Equal or similar group variances and roughly balanced group sizes

When to use something else

• Unequal group variances; use the Games-Howell test
• Only comparisons against one control group; use Dunnett's test

Assumptions and requirements

• Same assumptions as the ANOVA: independence, normality and equal variances
• Continuous outcome
• Tukey-Kramer adjustment when group sizes differ

What the analysis reports

• Pairwise mean differences
• Adjusted p-values
• Simultaneous 95% confidence intervals
• Homogeneous subsets or compact letter display

How to report it

Tukey HSD post hoc comparisons showed that [Group A] scored higher than [Group B], mean difference = [value], 95% CI [[value], [value]], p = [value].

Related methods

• One-Way ANOVA
• Games-Howell Test
• Bonferroni Correction

Run Tukey HSD Post Hoc Test on Quorant