Beta significance level in one model the same as the R-squared-change significance

#1
I am very new to statistics, so apologies in advance if I have not explained this clearly.

I have run two types of linear regression:
- The first contains 4 independent variables, all entered at the same time.
- The second uses "blocks" to enter 3 of the independent variables in the first block, and the 4th independent variable in the second block, in an attempt to calculate the unique variance explained by the 4th independent variable.

For three out of my four independent variables, when running the linear regression using blocks, the significance level for the R-squared-change is the same p value as the beta significance level in the original model (where everything was entered at once). For one of my independent variables, the two significance levels are different.

Why might this be? And does it indicate that I am doing something wrong...? :confused:

Thank you!
 

hlsmith

Omega Contributor
#2
Re: Beta significance level in one model the same as the R-squared-change significanc

To better address your agenda, there is an omega-square or eta-square value that you can calculate from the full model. This estimate tells you the partial R^2 contribution of each variable in the full model. I would recommend using these values for partial R^2 along with their 95% CIs. Much more straightforward and easier to explain.