ANOVA-Type Tests - Comparing Three or More Groups
Jump to navigation Jump to search
|Related Online Training modules|
|Planned Tests Of Statistical Significance|
|Generally it is best to access online training from within Q by selecting Help > Online Training|
ANOVA-type tests involve comparing three or more cells, most commonly in rows of a table. They are conducted automatically when:
- ANOVA-Type Test is selected in in the Column comparisons settings of Statistical Assumptions.
- When appropriate data is selected when pressing (i.e., when conducting Planned Tests Of Statistical Significance). See Planned ANOVA-Type Tests.
- When using Smart Tables with a Number question by a Pick One question.
There are two basic variants of these tests:
- Independent samples tests, which compare data between different sub-groups. This is sometimes referred to as One-Way ANOVA and One-Way Layouts. In Q, these occur when you have
- The table has numeric data in the rows (i.e., Number or Number - Multi question in the brown drop-down).
- The table has mutually exclusive categories within spans in the columns (i.e., has no NETs).
- Repeated measures tests, where each respondent has provided multiple responses and these are to be compared (e.g., evaluations of different products). These are sometimes referred to as Two-Way ANOVA and Two-Way Layouts. In Q, these occur when you have a Pick One – Multi question selected in the blue drop-down menu with the variables appearing as the columns. (Where you wish to compare averages rather than categories, select Statistics – Below and Average.)
Additionally, whether the setting for Means in Statistical tests for categorical and numeric data is set to Non-parametric and the Question Type of the question in the blue drop-down are also determinants of how the tests are computed, as described in the table below.
|Independent samples||Dependent samples|
|Not Non-parametric and the data in the rows is numeric (i.e., Number or Number - Multi)||F-Test (ANOVA)||Repeated Measures ANOVA with Greenhouse & Geisser Epsilon Correction .|
|Non-parametric and the data in the rows is numeric (i.e., Number or Number - Multi)||Kruskal-Wallis Test||Friedman Test for Correlated Samples|
|Non-parametric and the data in the rows is categorical (Pick One, Pick Any or Pick Any - Compact)||Pearson's Chi-Square Test of Independence||Cochran's Q|
|Non-parametric and the data in the rows is mutually exclusive (e.g., choices on a rating scale)||Pearson's Chi-Square Test of Independence||Chi-Square Test for Compatibility of K Counts|