'Assume equal variance' in contrasts? (JASP)
Hello!
When conducting polynomial contrasts with JASP, there is a possibility to untick an "Assume equal variance" box if your levels do not have the same variance.
My question is: what does unticking this option do ? What correction does it apply ?
Thank you!
Comments
Hi Da5Ax,
When you untick this option, you basically conduct a series of paired samples t-tests between the different contrast groups. These t-tests each use a separate error variance, rather than a single error variance (which is the default). Note that this can only be done for main effects, and not interaction effects.
Kind regards,
Johnny
Thank you!