Analysis of variance (ANOVA) is usually done one way. This means that only one variable is used to group individuals. An example would be if patients are allocated to treatment A and B. We would then have one variable named “treatment”. This variable may have the value A or B. It might have been that we had three groups A, B and C respectively. Differences in outcome between groups would often be analysed using one way ANOVA. One way ANOVA for two groups (A and B) is the same as Student’s t-test for independent groups.

In a few studies we use more than one variable to group individuals. Assume that we allocate individuals to four different treatments described with one variable having the value A, B, C or D it would normally be analysed with one way ANOVA. If we have two different treatments and two different time intervals before treatment is initiated group allocation would be described using two variables namned “treatment” and “delay”. The first would have the possible values A and B while the latter may have the values “early” and “late”. In this case we would use two way ANOVA to make the group comparison.