Analyses

Mann–Whitney U test

Compare two independent groups without assuming normality.

Edit on GitHub

When to use it

Use Mann–Whitney U instead of an independent-samples t-test when the outcome is ordinal or the scores are skewed. It asks whether one group tends to rank higher than the other.

Assumptions

The two groups are independent. The outcome is at least ordinal. Tensr does not run a normality check for this test.

Running it in Tensr

Analyze → Nonparametric Tests → 2 Independent Samples. In chat: “Mann–Whitney U of score by method.”
Group column is the two groups. Value column is the score.
Confidence level starts at 0.95. There is no effect-size option on this request.

Options

Prop

Type

Reading the output

Lecture versus workshop ranks on the exam. The workshop scores tend to sit higher, and the two lists still overlap, so U is not zero.

GroupnMedianMean rankRank sumStatus
Lecture3269.126.312842included
Workshop3272.9538.6881,238included

Significant difference in score between Workshop and Lecture (U = 314, p = .008, r = 0.332). Effect size (medium; r < 0.1 negligible, 0.1 ≤ r < 0.3 small, 0.3 ≤ r < 0.5 medium, r ≥ 0.5 large). Cutoffs: Cohen, 1988. The primary result is significant (p = .008). This is large enough, in this sample, to treat the comparison this page is about as a real association rather than noise. Metrics: U statistic = 314; p-value = .008; Effect size r = 0.332; Effect size = medium.

Reporting (APA 7)

Significant difference in score between Workshop and Lecture (U = 314, p = .008, r = 0.332). Effect size (medium; r < 0.1 negligible, 0.1 ≤ r < 0.3 small, 0.3 ≤ r < 0.5 medium, r ≥ 0.5 large). Cutoffs: Cohen, 1988. This result is significant (p = .008). The effect is moderate, so report its size with the p-value.

Coming from SPSS

Analyze → Nonparametric Tests → 2 Independent Samples. SPSS often prints Wilcoxon W and a z alongside U.