Quade Test

The Quade test is a nonparametric test for randomized complete block designs that weights each block's within-block ranks by the rank of the block's range and compares treatments with an F statistic.

Quade 검정은 완전 무작위 블록 설계에서 각 블록 안의 순위에 블록 범위의 순위로 가중을 주고 F 통계량으로 처리 간 차이를 비교하는 비모수 검정입니다.

I have several treatments measured on the same subjects; is there a nonparametric test more powerful than Friedman's for a few treatments?

같은 대상에게 처리 몇 가지를 모두 적용했는데, 처리 수가 적을 때 Friedman 보다 검정력이 높은 비모수 검정이 있나요?

The Quade test ranks the observations within each block, ranks the blocks by their range, and multiplies the centered within-block ranks by the block's range rank so that blocks with clearer treatment differences count more. The resulting statistic follows an F distribution with k - 1 and (b - 1)(k - 1) degrees of freedom.

When to use Quade Test

• Each subject or block receives all treatments and the outcome is ordinal or non-normal.
• There are few treatments (about three to five), where the Quade test tends to be more powerful than Friedman's.
• The values within a block can be compared on a common scale so that block ranges are meaningful.

When to use something else

• There are many treatments; the Friedman test is usually more powerful.
• The outcome is approximately normal; use repeated-measures ANOVA.

Assumptions and requirements

• Blocks are independent of each other.
• Observations within a block are comparable on the same scale, since ranges across blocks are ranked.
• There are no missing values within a block.

What the analysis reports

• Quade F statistic with degrees of freedom and p value
• Sums of squares A and B of the weighted rank scores
• Weighted rank sums, mean ranks and medians per treatment

How to report it

A Quade test showed a significant difference among the [k] treatments across [b] blocks, F([df1], [df2]) = [value], p = [value].

Related methods

• Friedman Test
• Conover Post Hoc Test
• Randomized Complete Block Design Analysis

Run Quade Test on Quorant