Hi everyone,
I have a base model and an alternative model. The R2 in the alternative model is greater than the R2 in the base model. It means the alternative model has greater explantory power. My question is, which test should i run to test the differences of these two models?
one more question: say i got two latent variables (LV1 and LV2)and they both link to this exogenous variable. LV1 is only significant when LV2 connects to the exogenous variable. I don't understand why is this. Can anyone help me with this?
I hope i make my questions clear. Thank you very much for any help!
Best Regards,
Catherine
compare models
Hi all,
concerning Catherines first question: I have the same problem. In my case, the alternative model is an extension of the base model. Can we conduct an ordinary F-test to test for R² differences (e.g. as described here: http://psych.unl.edu/psycrs/statpage/rhtest_eg1.pdf)?
Best,
Ralf
concerning Catherines first question: I have the same problem. In my case, the alternative model is an extension of the base model. Can we conduct an ordinary F-test to test for R² differences (e.g. as described here: http://psych.unl.edu/psycrs/statpage/rhtest_eg1.pdf)?
Best,
Ralf
Hi Ralf,
I am not sure if you still read this forum.
Here is the formula how Chin(2010) uses F-test to examine the R² difference:
F=(R₂²- R₁²/K₂-K₁)/(1-R₂²/N-K₂-1)
with K₂-K₁ and N-K₂-1 degree of freedom
Where R₁² is for the base model and R₂² is the superset model that includes the additional Latent Variables, K₂ is the number of predictors for the super model and K₁ is the number of predictors for the baseline, and N is the sample size.
It is in Chapter 28 of Handbook of PLS written by Chin (2010).
Best regards,
Catherine
I am not sure if you still read this forum.
Here is the formula how Chin(2010) uses F-test to examine the R² difference:
F=(R₂²- R₁²/K₂-K₁)/(1-R₂²/N-K₂-1)
with K₂-K₁ and N-K₂-1 degree of freedom
Where R₁² is for the base model and R₂² is the superset model that includes the additional Latent Variables, K₂ is the number of predictors for the super model and K₁ is the number of predictors for the baseline, and N is the sample size.
It is in Chapter 28 of Handbook of PLS written by Chin (2010).
Best regards,
Catherine