Presumably by accident, this summary omits a comment
included in my own reply about Stata:
In the case of Stata, your comment is misrepresentation
insofar as it might lead to the impression
that Stata in some sense encourages this
kind of model. You don't name the command,
but this is not a fair summary of -xi- if that
is what you have in mind. Stata neither encourages
nor discourages this practice: the user decides
what model to fit.
Nick
[log in to unmask]
Jon Heron (ALSPAC)
> thanks for all the replies to my recent email.
> Please find below a repeat of the original query, and a list
> of anonymised
> replies.
|