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 on April 23, and its replacement, statalist.org is already up and running.


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

Re: st: interpretation for negative and positive slope combination of interaction term


From   David Hoaglin <dchoaglin@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: interpretation for negative and positive slope combination of interaction term
Date   Thu, 9 May 2013 18:06:06 -0400

Dear Nahla,

You are correct about the slope and intercept, as shown in Figure 7.8
of that book.  In David Crow's simplified model, B2 does not
contribute to the slope for MV.

A major part of your difficulty comes from trying to make ratio
interpretations of coefficients in a model whose coefficients can tell
you about differences.  One usual way to get coefficients that have a
straightforward interpretation as ratios is to analyze the dependent
variable in a logarithmic scale.  If it is not appropriate to work
with your dependent variable in a log scale, you may be able to use a
generalized linear model with a log link function.

David Hoaglin

On Thu, May 9, 2013 at 3:51 PM, Nahla Betelmal <nahlaib@gmail.com> wrote:
> Thanks for the reply David, but I think there is something not quite
> right. If you check this file, p.133 figure 7.8
>
> http://www.sagepub.com/upm-data/21120_Chapter_7.pdf
>
> You will notice that in order to get the slope for the group with
> Dummy= 1 (overconfident manager in my case), we should add the
> coefficient of beta and gama ( MV and OC*MV in my case) , and to get
> the intercept for those managers we should add the alpha and y ( my
> model intercept and the coefficient of OC )
>
>  I found the same in other files, the problem is that all the examples
> they provide both beta and gama are positive which makes the addition
> and interpretation process easy.
>
>  my question is how to add and interpret when one is positive and the
> other is negative . As the effect of MV on realistic managers is -
> 0.0566  and on  overconfident managers is + 0.0596 , it moves from
> negative to positive 0.003 (again the interaction is significant
> although at 10 level). how many times MV effect overconfident managers
> more than other managers ?
>
> I would really appreciate help in that
>
> Many thanks
>
> Nahla
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


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