jamovi

Hello:
I have a suggestion here for the linear regression module. The module allows us to include interaction effects. When we do interaction effects, we often mean center the variables before we calculated their interaction term. Is that possible to include this option in the module, so that we do not need to manually do the mean center for the analysis?
Thanks

you can use GAMLj for that. It centers the variables by default. linear regression module does the standard regression, so centering would be confusing

mcfanda@gmail.com wrote:you can use GAMLj for that. It centers the variables by default. linear regression module does the standard regression, so centering would be confusing

but GAMLj lacks another function: it does report change of R square. The linear regression model does that, but do not provide mean center. So neither of them can satisfy the need.

it's pretty easy to center variables using the computed/transformed variable system.

i.e. varName - VMEAN(varName)

cheers