The normal thing to conclude, is to say that you have gotten indirect support for the alt hyp that IE affects DL (since the association was significant), but that this was not the case for the alt hyp that PC affects DL (since the association was not significant). That is, one of your hyps got support :-)

It's normal and common for one or more of your variables to not be a significant predictor of the outcome variable in an otherwise significant omnibus model.

Given that your IE is significant but PC is not, your discussion section cannot conclude that PC still influences DL (this would be misinterpreting the results).

Your conclusion should rather be that DL is significantly predicted by IE controlling for PC, but is not predicted by PC controlling for IE. I personally do not give much attention to the significance of the omnibus model because the overall performance of the model is rarely ever the interesting part. Rather, you can just interpret the R^2 value by stating that the examined model explains __% of the variance of DL.

Ao Li

Hi Ao Li. You can interpret in this way that out of two predictors IE is a significant predictor of DL and it explains about X percent (write the X based on the Beta value) of the variation in the DL. In this case, even if one IV is significant and one is not, you can indicate that the null hypothesis is rejected.