If you specify a continuous Y and a nominal/ordinal X in Bivariate Relationships, a one-way report is displayed.
This report provides various analysis options via the red triangle in the upper left.
[Mean/ANOVA/Pooled t-test]
Perform a t-test assuming equal variances.
[t-test using individual variances]
Performs a t-test under the assumption of unequal variances, also known as Welch's test.
[Compare Means] → [Each Pair, Student's t-test]
We use a method called the unrestricted LSD method, which calculates the error variance from all groups to calculate the test statistic.
Therefore, when comparing means for data with three or more groups, the test results output in the Difference Rank Report will be different from the test results output in Means/ANOVA/Pooled t-Test when the data is limited to only two groups.
