if I fit the regression analysis for a nominal variable (manipulated as two conditions) as IV, and a continuous variable as DV, at the same time, I add three covariates into the model. I wonder whether the covariates would be automatically mean-centred in the result table for this regression. for example, the "parameter estimates table" shows:
"intercept"
"nominal variable"
"covariate 1"
"covariate 2"
"covariate 3"
moreover, once I add an interaction effect between the nominal variable and another continuous variable. I awarded that the continuous variable is mean-centred. For example: the "parameter estimates table" table shows
"intercept"
"nominal variable"
"covariate 1"
"covariate 2"
"covariate 3"
"another continuous variable"
"nominal variable * another continuous variable (average- 3.13)"
I know 3.13 is the mean value of "another continuous variable", but I am not sure whether "covariate1", "covariate 2" "covariate 3" and "another continuous variable" are mean-centred or not in this multiple regression model.
Hi @Rongyu_Kuang : No, the others are not mean centered. You can see this via the "Show Prediction Expression" option in the menu (red triangle) next to "Response" in the output. See pic below.
Please see the documentation for the linear model terms used in Fit Least Squares.