I need to be able to critically assess a regression analysis printout from EViews (sample attached) and be able to identify possible issues - i.e.:
- potential heteroskedasticity
- potential autocorrelation
- potential multicollinearity problem
prior to running the specific tool which provides further analysis for one of these 3 issues. Also, assessing the f-statistic and obs*r-squared from one of these 3 tests to confirm or reject the existence of a problem.
Knowing what values in what fields will indicate statistical significance is essentially what I need.
NOTE: The attached file is a sample printout (the data is not correct) - I need to know 'in general' what to look out for in a printout.
I think what you are generally asking is how to interpret output from a hypothesis test whether it is for heteroskedasticity, autocorrelation, or multicollinearity. When you run a hypothesis test the null is that the condition doesn't exist or that there is no heteroskedasticity, autocorrelation, or multicollinearity. The test statistic is used to try to reject ...
The solution provided is a 200-250 word general explanation of how to interpret whether output from a hypothesis test is for heteroskedasticity, autocorrelation, or multicollinearity and three .pdf attachments which got into more depth on the topic.