Try writing your model for some specific values of age and exp
You would have
Ln(E(injury))=B0+B1*25+B2*exp+B3*25*exp+ln(hrs_day) for age 25
Ln(E(injury))= B0+B1*30+B2*exp+B3*30*exp+ln(hrs_day) for age 30
The difference is 5*B1+5*exp*B3
Testing B3=0 is a test for same slope of regression line.
The intercept is the same if B1 is 0 and B3 is 0. It may be helpful for
interpretation to either center at the means, or find a 'smallest useful
age' at center there. For example, use age 20 if that's a useful
minimum age. Then the intercept B0 will refer to values at age 20.
Tony
Peter A. Lachenbruch
Department of Public Health
Oregon State University
Corvallis, OR 97330
Phone: 541-737-3832
FAX: 541-737-4001
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of
[email protected]
Sent: Tuesday, January 20, 2009 3:12 AM
To: [email protected]
Subject: st: Testing for Significance Between Poisson Reg Lines
I was wondering how I can go about testing for a statistically
significant difference between the two lines produced below?
age=age of person
exp=amount of time doing a certain job
hrs_day=number of hours worked on a given day
injury=whether or not the person was injured (0/1)
hrs_day and injury are summed by age and exp. age and exp are integers.
poisson injury age exp ae if hrs_day>0, exposure(hrs_day)
replace age= 25
replace ae= 25*exp
predict y_age25, ir
replace age = 30
replace ae = 30*exp
predict y_age30, ir
twoway (line y_age25 exp)(line y_age30 exp)
Thanks.
*
* 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/
*
* 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/
