Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: Probabilities as predictor variables


From   Nick Cox <njcoxstata@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Probabilities as predictor variables
Date   Wed, 18 Apr 2012 16:58:47 +0100

A second take is that I don't see any objection to a predictor defined
on the interval [0,1] or (0, 1). But what functional form would make
sense? Depends on the application, but it might well be that you would
be better off working with say logit x rather than x, for similar but
not identical reasons to those for using logits for the response in
the usual model.

Nick

On Wed, Apr 18, 2012 at 4:22 PM, Nick Cox <njcoxstata@gmail.com> wrote:
> I am not sure what distinction you are making here. But if a predictor
> is truly binary, then there can only be two predictions. Can still be
> useful: it's classically a t-test comparing two means and in most
> people's books just another regression.
>
> On Wed, Apr 18, 2012 at 4:00 PM, Ronald McDowell
> <McDowell-R3@email.ulster.ac.uk> wrote:
>> Hi
>>
>> I am able to fit a model in which a binary variable (0/1) is the dependent variable and another variable, such as age, is the predictor variable using logistic regression.
>>
>> Is it possible in STATA to fit a model in which these roles are reversed ie the binary variable is the independent variable (treated as a probability rather than as a dummy variable) and age is the outcome?
>>
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index