How to Replicate SPSS Significance Tests in Q
Although there is a relatively large overlap between the statistical tests in SPSS Statistics and Q, very few of the default tests in Q match the default tests available in SPSS. If seeking to compare test outputs between the programs the following needs to be kept in mind:
- In Q, test selection is usually done by changing the data setup, rather than by selecting specific tests (see Tests Of Statistical Significance).
- Most statistical tests in SPSS assume that either the data is unweighted, or, that any weights are frequency weights. Q, by contrast, assumes that weights are sampling weights. See Using Frequency Weights in Q for instructions for making Q treat the weights as frequency weights.
- The rule called Significance Testing - Independent Samples Column Means and Proportions Tests has been designed to replicate some commonly used SPSS tests.