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: Strange Behaviour When Selecting Levels For Factor Variables In Regression With i#
From
Richard Williams <[email protected]>
To
[email protected], [email protected]
Subject
Re: st: Strange Behaviour When Selecting Levels For Factor Variables In Regression With i#
Date
Sat, 19 Jan 2013 01:57:23 -0500
You understand fine (except it wouldn't matter even if sex had more
than 2 categories -- changing the reference category used will only
change the coefficients for sex and the constant, not the other
variables in the model). There is some sort of problem on Dan's end.
At 01:12 AM 1/19/2013, Sarah Elizabeth Edgington wrote:
I share Daniel's confusion and this explanation doesn't make it any
clearer to me.
If sex were not a binary variable then the other coefficients
changing would make sense because the different regressions would
actually be changing how sex is coded. However, with a variable
that's coded 0/1 to begin with, isn't i.sex the equivalent of
i1.sex? That is, isn't an indicator for when sex=1 exactly the same
variable as the original sex variable? In which case you'd expect
the first two regression examples to be exactly the same. The fact
that they aren't suggests to me that there's something I don't
understand about the i1.sex factor variable syntax. What am I missing?
-Sarah
At 08:32 PM 1/18/2013, you wrote:
Daniel,
The definition of a coefficient in a multiple regression model
includes the list of all the other predictors in the model. Thus,
using a different predictor for sex changes the definitions of the
coefficients for patient and the categories of when.
If the other predictors were orthogonal to sex, the numerical values
of their coefficients would not differ among the models, even though
the definitions were not the same.
Your models may be equivalent, in the sense that they have the same
predicted values and the same residuals, but they are not all the same
model.
I hope this discussion helps.
David Hoaglin
On Fri, Jan 18, 2013 at 6:45 PM, <[email protected]> wrote:
> Hello,
>
> when i use indicator i with selecting level of a factor variable
like i1.varname to run a regression I get strange results.
>
> For example:
>
> sysuse blong,clear
> regress bp i.sex i.when c.patient i.when#c.patient
> regress bp i1.sex i.when c.patient i.when#c.patient
> regress bp i0.sex i.when c.patient i.when#c.patient
>
> This regression is wihout sense but theoretically it should
estimate the same model and should give same results except for
variable sex cause all I do is demand an indicator for a different
level of a 2-level variable sex.
> But if I run these lines I get three regressions with three
different coefficients for the variable "when" and "patient" even
I didnt change anything that should be related to these variables.
> Whats wrong here?
>
> regards
> Daniel
*
* 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/
*
* 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/
-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
HOME: (574)289-5227
EMAIL: [email protected]
WWW: http://www.nd.edu/~rwilliam
*
* 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/