[Gretl-users] Casewise Diagnostic

[Sven Schreiber]

Novan AS =^-^= schrieb:
> Dear All,
>
> I want to request for a casewise diagnostic in the OLS regression.
> The Casewise Diagnostic is to find the outlier in a regression.
> It's can be easily find with SPSS program.
> However in SPSS, we have to manualy delete the observation that found
> to be outlier.
> So, can gretl make an option before running a regression. 
> User can click an option (with any number of standard deviation) that
> automatically not to include the outlier observation to the regression.
> However the result have to inform the observation that become outlier.
>
>   
Hi Novan,

I'm not the one to decide this but for time series regressions I'm not 
sure it's good to make it automatic. There are subtle issues with 
outlier removals in dynamic contexts and especially when you have 
possibly integrated time series. I guess in cross-section regressions 
(SPSS's main domain, right?) it's not that much of a problem.

Having said that, in PcGive there is a post-estimation feature that 
reports observations with large residuals (per default I think it's 3.5 
standard devs, but it's adjustable). I have used that feature many times 
and maybe it's something to think about for gretl (after 1.6.6 of course).

cheers,
sven