Fligner-Killeen Test

The Fligner-Killeen test is a rank-based test of homogeneity of variances that converts absolute deviations from group medians into normal scores and compares their group means.

Fligner-Killeen 검정은 집단 중앙값에서의 절대편차를 순위화한 정규 점수로 바꿔 집단 간 평균을 비교하는, 순위 기반 등분산성 검정이다.

My data are not normal; which test should I use to check whether group variances are equal?

자료가 정규분포가 아닌데 집단 분산이 같은지 확인하려면 어떤 검정을 써야 하나요?

The Fligner-Killeen test ranks the absolute deviations of observations from their group medians and transforms the ranks into normal quantile scores. Because it uses ranks and medians, it is among the most robust tests of equal variances when data depart from normality.

When to use Fligner-Killeen Test

• Equality of variances must be checked before an ANOVA or t test.
• The outcome is skewed or has outliers.
• Spread itself is the research question, such as comparing variability between processes.

When to use something else

• Data are clearly normal and only two groups are compared; the F test for two variances has more power.
• Means rather than spreads are of interest; use Welch's ANOVA.

Assumptions and requirements

• Observations are independent.
• The outcome is at least ordinal and continuous enough to rank.
• Groups share a common distribution shape apart from scale.

What the analysis reports

• Chi-square statistic
• Degrees of freedom (groups minus one)
• p value
• Group medians, variances and median absolute deviations

How to report it

A Fligner-Killeen test indicated that variances were not homogeneous across groups, χ²([df]) = [chi2], p = [p].

Related methods

• Levene's Test
• Bartlett's Test
• F Test for Equality of Two Variances

Run Fligner-Killeen Test on Quorant