EViews.com

Posted: **Tue Apr 07, 2015 7:48 am**

Hey,

We could just use robust standard errors, but if devising an asymptotically efficient linear probability model (using WLS), one should first ensure fitted values are not greater than 1 or less than zero.

I have coded a transformation for an adjustment, but it takes two entries:

Code: Select all

series yhat = @recode(fitted>1,.999,fitted)

Code: Select all

series yhat2 = @recode(yhat<0,.001,yhat)

is there a way to combine these adjustments into a stephen hawking worthy adjustment for everything?

Thanks

Posted: **Tue Apr 07, 2015 8:43 am**

I don't understand a word of your question, but I think you're asking if you can combine those two series statements into one:

Code: Select all


series yhat2 = @recode(@recode(fitted>1,.999,fitted)<0,.001,@recode(fitted>1,.999,fitted))

Posted: **Tue Apr 07, 2015 9:05 am**

yes, thank you. Sorry for the confusion.

EViews.com

Unit Interval Adjustment

Unit Interval Adjustment

Re: Unit Interval Adjustment

Re: Unit Interval Adjustment