Statalist


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

st: RE: stata graphics- qfitci versus lfitci??


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: stata graphics- qfitci versus lfitci??
Date   Fri, 29 May 2009 10:51:00 +0100

The question may be quick, but what it means is not clear to me. 

In what sense can the -lfit-s be better than the -qfit-s? If you fit a
quadratic to a linear relationship, that should be fine because the
linear's just a special case. Sure, you are using two parameters rather
than one, so there's an issue of parsimony, but what Allison means by
"better" here I don't know. 

Otherwise, yes: you can mix two different kinds of graphs if you think
that's a good idea, but why is there any doubt about that? Does "can"
mean "should"? 

Nick 
n.j.cox@durham.ac.uk 

Allison.Milner@student.griffith.edu.au

I have a quick question regarding the graphs. I am trying to graph the
relationship between some interaction variables and my outcome variable.
To do this, I am using "twoway (qfitci DV IV)". However, some of the
relationships are better represented using 'twoway (lfitci DV IV)'. Can
I use a mixture of both of these approaches to present my data?

*
*   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   |   What's new   |   Site index