Statistics Glossary → Between-subjects factor
Between-subjects factorconcept
A factor in which each subject appears at only one level, so that its levels separate different groups of subjects (for example sex or treatment group).
Factors in an experimental design come in two kinds. In a between-subjects factor each subject is assigned to, or belongs to, a single condition, so the comparison is between different people. In a within-subjects (repeated) factor the same subject is measured under every condition, so the comparison is between the same person's measurements. For example, 'drug or placebo' is a between-subjects factor, while 'before treatment, month 1 and month 3' is a within-subjects factor.
Between-subjects effects are tested against an error term that includes the natural differences between people, so detecting the same effect usually needs more subjects than in a within-subjects design. In return, order and carryover effects are not a problem. Designs that combine the two kinds of factor are called mixed (split-plot) designs, and their most interesting result is often the group x time interaction.
An everyday example
In a sleep study, half of the participants use a screen before bed and half do not; each person is in only one group, so 'screen use' is a between-subjects factor. If sleep quality is measured in the same people in weeks one, two and three, 'week' is a within-subjects factor. The key question is whether sleep quality changes differently over the weeks in the two groups.
In MerQur
In Repeated Measures ANOVA it is added through the 'Between-subjects factor(s) and Covariate(s)' field; Mixed-Design ANOVA combines the between-subjects (group) factor with a within-subjects factor and can take a second one through the 'Additional within-subject factor, additional between-subjects factor, and covariate(s)' field. The 'Second factor and covariates' option in MANOVA and the '2nd factor' option in ANCOVA add a second between-subjects factor.
Analyses where this concept appears
Repeated Measures ANOVAMixed-Design ANOVA, Split-PlotMANOVAANCOVA