test difference between two R2 of nested models through the Bias corrected accelerated confidence intervals

Frequently asked questions about PLS path modeling.
Post Reply
ElNicolosi
PLS Junior User
Posts: 3
Joined: Tue Jul 02, 2019 8:18 am
Real name and title: Dr. Eleonora Nicolosi

test difference between two R2 of nested models through the Bias corrected accelerated confidence intervals

Post by ElNicolosi » Tue Jul 02, 2019 8:34 am

Hello,

Hope you are all well!

I have a problem and I really hope you can help me.

I would like to test if the difference in the R2 of two nested models is significant. For both models, I am using exactly the same data-set. In the second model, I just added an exogenous construct. Variables are not normally distributed.

I got Bias-Corrected accelerated confidence intervals for R2 values in both models through bootstrapping procedure (5000 samples). My questions are:

1. It is safe to say that if those confidence intervals do not overlap then the two R2 are different? Or should I account for the number of variables added before comparing the R2? how? Maybe I could use confidence intervals fro R2 adjusted?

2. Also, I was thinking to just get the differences of R2 for each sample produced through the bootstrap procedure. Then construct the Bias-corrected confidence interval on the distribution of those differences. Would that be ok?

3. If I need to compare R2 on more than two nested models should I build confidence intervals accounting for the Bonferroni correction?

Do you have any suggestions?

Thank you for reading the above!
E.

Post Reply