Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


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

re: st: effect coding with interaction variables


From   "Airey, David C" <[email protected]>
To   "[email protected]" <[email protected]>
Subject   re: st: effect coding with interaction variables
Date   Wed, 10 Mar 2010 06:57:13 -0600

.

Phil Ender has written "anovalator" which gets you interaction tests from indicator coded regression _as if_ you had used effect coding (http://www.ats.ucla.edu/stat/stata/faq/anovalator.htm). Especially, see the FAQ on use of the margins command here, if you are using Stata 11 (http://www.ats.ucla.edu/stat/stata/faq/default.htm).

You can also get what you could have gotten if you used effect coding with indicator coding by hand using test after regression but it seems more complicated than if you had just used effect coding in the first place (http://www.ats.ucla.edu/stat/stata/faq/main_effects.htm).

If you want to learn about manually coding different coding schemes, there is some material online here (http://www.ats.ucla.edu/stat/stata/webbooks/reg/chapter5/statareg5.htm).




> Hi
> 
> I would like to get some information about effect coding in regression
> especially with interaction variables. Any reference or example will be
> highly appreciated.
> Just to give you an idea about the data which I am using:
> 
> I have 5 Road classes, 5 vehicle classes, 24 interaction variables. In
> effect coding I am using 4 road classes, 4 vehicle classes and 15
> interaction variables. I am coding them as ( 1 if it is comparison group,
> 0 if it belongs to other 2 groups, -1 if is the 5th group which I am
> omitting  as explanatory variable. I am doing the same stuff with 15
> interaction variables (1 if it is comparison group, 0 if it belongs to
> other 14 groups, -1 if is the reference group which I am omitting as
> explanatory variable.
> 
> If some can comment or direct me to any relevant example/ literature I
> will be highly grateful.
> I have also used Devcon but it seems it is not going to work in this case
> as in interactions it is having more than one reference category and it is
> predicting different values than the original model.
> 
> Regards
> 
> Memon

*
*   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 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index