A fractional logit is likely appropriate.
Here is a reference:
Econometric Methods for Fractional Response Variables With an Application to
401 (K) Plan Participation Rates, Leslie E. Papke and Jeffrey M. Wooldridge
Journal of Applied Econometrics, Vol. 11, No. 6 (Nov. - Dec., 1996), pp.
619-632
Here is the code in stata with robust standard errors.
glm y x1 x2 ... xk, fam(bin) link(logit) robust
----- Original Message -----
From: "Ashwin Ananthakrishnan" <[email protected]>
To: <[email protected]>
Sent: Tuesday, August 04, 2009 7:56 PM
Subject: st: Poisson vs. Linear regression for comparing rates
Hi,
I'm trying to do a state-level analysis looking at rates of a particular
cancer and it's relationship to some state-level variables such as
proportion uninsured in the state, racial composition of the state and
some other state-level economic variables.
However, I'm not sure if I should be using Poisson regression or Linear
regression to assess the relationship between the rate of cancer
(independent variable) and the dependent variables.
What determines which regression model is the best?
Should I automatically start with Poisson regression, and use the goodness
of fit test after that to see if that's the appropriate model (i.e. if gof
p > 0.05). If so, should I do it after univariate analysis with each
variable?
If the rate of cancer is not normally distributed, should I log transform
it and then use linear regression?
Thanks in advance for your help.
Ashwin
*
* 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/
--------------------------------------------------------------------------------
No virus found in this incoming message.
Checked by AVG - www.avg.com
Version: 8.5.392 / Virus Database: 270.13.43/2281 - Release Date: 08/04/09
05:57:00
*
* 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/
