Page 1 of 1
Significance testing in eview
Posted: Mon Nov 09, 2015 1:28 pm
by xchuj1
Is there a chance to get some detail on the test of significance used in the cross-correlation function in eviews (not disucssed in the eviews manual)? The cross-correlogram provides the actual (cross) correlation coefficients and indicates the significance only graphically with asterisks. I need to get some detail on how these asterisks are determined. Thank you.
Re: Significance testing in eview
Posted: Mon Nov 09, 2015 2:02 pm
by EViews Gareth
Could you post an example of the significance you are interested in?
Re: Significance testing in eview
Posted: Mon Nov 09, 2015 2:10 pm
by xchuj1
Please see the attached correlogram from the eviews manual. The balk charts on the left hand side (which in another format are replaced by asterisks) should represent the significance of the
correlation coefficients (thats at least the way I understand it). I would need to do it a bit more sophisticated and actually calculate the significance of the correlation coefficients. In order to do that I need to figure out what test of significance eviews uses. Thanks for your help.
Re: Significance testing in eview
Posted: Mon Nov 09, 2015 2:12 pm
by EViews Gareth
There is no significance reported. Just the size of the coefficients
Re: Significance testing in eview
Posted: Mon Nov 09, 2015 2:21 pm
by xchuj1
Understood. That was my misunderstanding then. This however only addressed my problem partially. Can you pls advise me how to (with or w/o eviews) test the significance of the correlation coefficients from this cross-correlogram...
Re: Significance testing in eview
Posted: Mon Nov 09, 2015 2:22 pm
by EViews Gareth
That's an econometric issue to which I do not know the answer