While learning multiple regression, I once meet the concepts of “complete confounding” and “partial confounding”

In the model, the response variable is referred to as Y. There have two predictor variables: X1 and X2. With respect to the complete confounding, the correlation between X1 and X2 is 1.

With respect to the partial confounding, the correlation between X1 and X2 is 0.95237.

I attached the ANOVA tables for both cases. The first one is for complete confounding and the second one is for partial confounding.

I noticed that for complete confounding case, some F-statistics and P-value are just N/A in the corresponding ANOVA table, why?

Secondly, looks like the P-value in the ANOVA table corresponding to the partial confounding is way smaller than the one related to the complete confounding case, why?

In the model, the response variable is referred to as Y. There have two predictor variables: X1 and X2. With respect to the complete confounding, the correlation between X1 and X2 is 1.

With respect to the partial confounding, the correlation between X1 and X2 is 0.95237.

I attached the ANOVA tables for both cases. The first one is for complete confounding and the second one is for partial confounding.

I noticed that for complete confounding case, some F-statistics and P-value are just N/A in the corresponding ANOVA table, why?

Secondly, looks like the P-value in the ANOVA table corresponding to the partial confounding is way smaller than the one related to the complete confounding case, why?

Last edited: