Odd thing happening in linear regression when I enter my predictors...can you help?

[sampier85]

Hello friends

This may sound like a total rookie question but please if you could help me better understand this. I am performing a linear regression with 2 predictors. When I run a linear regression using only one of these predictors, they are both significant in the model.

But, when I run a multiple regression putting both these predictors together, they both lose significance?

What is happening...how can I interpret this? Could this indicate mediation?

Thank you

DV: Depression
IV1: Dependency
IV 2: Neuroticism

 

[espenmei]

Re: Odd thing happening in linear regression when I enter my predictors...can you hel

Hi,

The variables (predictors) are probably correlated.

A solution might be to combine them to a single variable

 

[noetsi]

Re: Odd thing happening in linear regression when I enter my predictors...can you hel

Several things can cause this. One of the most common is high multicolinearity (which may not always show up as corelation although with just two predictors it should). You should ask for a VIF or tolerance statistic and see if this is occuring. Or look at the correlation between the predictors. It will probably be very high.

 

[Dason]

Re: Odd thing happening in linear regression when I enter my predictors...can you hel

Correlation between the predictors is probably the culprit (as the others have mentioned) but what is your sample size in this case?