Hi,

I hope that someone can give me a pointer please. I am trying to determine which is the most appropriate test to apply to my data.

I have some data which were originally formulated for a 2x2 Repeated Measures ANOVA. I was particularly interested in the A*B interaction. However, having computed the relevant analyses, it has become clear to me that the intervention applied to modulate factors A and B is dependent on a fixed interval covariate, so I need to change my approach.

I have initially run this as an ANCOVA, but I think it actually makes more sense to reformat my data by collapsing each factor to produce 2 single level factors, (so 2 DVs), with the covariate. I thus have one datum for each of the two factors, plus one for the covariate (so 3 values all together). What I am interested in, is whether A or B can be explained by the covariate. I am anticipating that A will, but not B. I should mention, the covariate is constrained to a value between 0 and 1.

Maybe I'm just over complicating this; I have run as two linear regression analyses, but I don't think that's quite right.

Can anyone help please?

Many thanks.

I hope that someone can give me a pointer please. I am trying to determine which is the most appropriate test to apply to my data.

I have some data which were originally formulated for a 2x2 Repeated Measures ANOVA. I was particularly interested in the A*B interaction. However, having computed the relevant analyses, it has become clear to me that the intervention applied to modulate factors A and B is dependent on a fixed interval covariate, so I need to change my approach.

I have initially run this as an ANCOVA, but I think it actually makes more sense to reformat my data by collapsing each factor to produce 2 single level factors, (so 2 DVs), with the covariate. I thus have one datum for each of the two factors, plus one for the covariate (so 3 values all together). What I am interested in, is whether A or B can be explained by the covariate. I am anticipating that A will, but not B. I should mention, the covariate is constrained to a value between 0 and 1.

Maybe I'm just over complicating this; I have run as two linear regression analyses, but I don't think that's quite right.

Can anyone help please?

Many thanks.

Last edited: