Stuart-Maxwell Marginal Homogeneity Test

The Stuart-Maxwell test assesses marginal homogeneity in a square contingency table of paired categorical ratings, testing whether the category distribution is the same on both occasions or raters.

Stuart-Maxwell 검정은 짝지은 범주형 평가의 정방 교차표에서 두 시점(또는 두 평가자)의 범주 분포가 같은지, 즉 주변 동질성을 검정한다.

The same respondents chose among three categories before and after an intervention; did the distribution of choices change?

같은 응답자가 개입 전후에 세 범주 중 하나를 골랐는데, 선택 분포가 바뀌었나요?

The Stuart-Maxwell test compares the row and column totals of a K × K paired table using a quadratic form in their differences with K − 1 degrees of freedom. With two categories it reduces to McNemar's test without continuity correction.

When to use Stuart-Maxwell Marginal Homogeneity Test

• The same units are classified twice into the same set of three or more categories.
• The question is whether the overall category distribution shifted.
• Two raters classify the same items and systematic disagreement in marginals is of interest.

When to use something else

• Only two categories are used; McNemar's test is the standard choice.
• The two classifications come from independent samples; use the chi-square test of independence.

Assumptions and requirements

• Pairs are independent of each other.
• Both classifications use the same categories.
• Enough discordant pairs exist for a chi-square approximation.

What the analysis reports

• Chi-square statistic
• Degrees of freedom
• p value
• Row and column marginal totals with their differences

How to report it

A Stuart-Maxwell test indicated that the marginal distributions of the two ratings differed significantly, χ²([df]) = [chi2], p = [p], N = [n].

Related methods

• McNemar's Test
• Cochran's Q Test
• Cohen's Kappa

Run Stuart-Maxwell Marginal Homogeneity Test on Quorant