Hello.
Please can someone help me? I ran the regression equation below with died being a binary outcome; the variable urban is a binary moderator (equals 1 if urban location and 0 if rural); period is a 5-year time period treated as a categorical variable with 5 levels; age is a continuous covariate and female is a binary covariate. The idea of this analysis is to investigate whether the trend in death rate is different for urban and rural locations. Various tests have indicated that the period variable should not be treated as a continuous predictor. Thus I will want to know whether the difference in mortality rate between period 1 and period 5 is similar for urban and rural locations.
My questions are:
1. What is the correct interpretation for the coefficients produced by this margins command
2. What code will I use to get the exponentiated simple slopes (separate slopes for urban and rural locations) of the mortality difference between period 1 and period 5, for example?
I will greatly appreciate any insight.
. svy, subpop(if analysispop == 1): glm died i.urban##i.period age female, family(poisson) link(log) eform
. margins urban, subpop(if analysispop == 1) dydx(period) vce(unconditional)
Average marginal effects Number of obs = 1,670,518
Subpop. no. obs = 157,603
Expression : Predicted mean died, predict()
dy/dx w.r.t. : 2.period 3.period 4.period 5.period
---------------------------------------------------------------------------------
| Linearized
| dy/dx Std. Err. t P>|t| [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.period | (base outcome)
----------------+----------------------------------------------------------------
2.period |
urban |
Rural | -.0219785 .0179504 -1.22 0.221 -.0571619 .0132049
Urban | -.0512376 .0061109 -8.38 0.000 -.0632151 -.0392601
----------------+----------------------------------------------------------------
3.period |
urban |
Rural | -.0243589 .0182387 -1.34 0.182 -.0601073 .0113895
Urban | -.0925881 .0059041 -15.68 0.000 -.1041603 -.0810159
----------------+----------------------------------------------------------------
4.period |
urban|
Rural | -.0545424 .0181749 -3.00 0.003 -.0901657 -.0189191
Urban | -.1336001 .0051976 -25.70 0.000 -.1437876 -.1234126
----------------+----------------------------------------------------------------
5.period |
urban |
Rural | -.1943786 .0164296 -11.83 0.000 -.226581 -.1621761
Urban| -.243518 .0056055 -43.44 0.000 -.2545049 -.2325311
---------------------------------------------------------------------------------
Please can someone help me? I ran the regression equation below with died being a binary outcome; the variable urban is a binary moderator (equals 1 if urban location and 0 if rural); period is a 5-year time period treated as a categorical variable with 5 levels; age is a continuous covariate and female is a binary covariate. The idea of this analysis is to investigate whether the trend in death rate is different for urban and rural locations. Various tests have indicated that the period variable should not be treated as a continuous predictor. Thus I will want to know whether the difference in mortality rate between period 1 and period 5 is similar for urban and rural locations.
My questions are:
1. What is the correct interpretation for the coefficients produced by this margins command
2. What code will I use to get the exponentiated simple slopes (separate slopes for urban and rural locations) of the mortality difference between period 1 and period 5, for example?
I will greatly appreciate any insight.
. svy, subpop(if analysispop == 1): glm died i.urban##i.period age female, family(poisson) link(log) eform
. margins urban, subpop(if analysispop == 1) dydx(period) vce(unconditional)
Average marginal effects Number of obs = 1,670,518
Subpop. no. obs = 157,603
Expression : Predicted mean died, predict()
dy/dx w.r.t. : 2.period 3.period 4.period 5.period
---------------------------------------------------------------------------------
| Linearized
| dy/dx Std. Err. t P>|t| [95% Conf. Interval]
----------------+----------------------------------------------------------------
1.period | (base outcome)
----------------+----------------------------------------------------------------
2.period |
urban |
Rural | -.0219785 .0179504 -1.22 0.221 -.0571619 .0132049
Urban | -.0512376 .0061109 -8.38 0.000 -.0632151 -.0392601
----------------+----------------------------------------------------------------
3.period |
urban |
Rural | -.0243589 .0182387 -1.34 0.182 -.0601073 .0113895
Urban | -.0925881 .0059041 -15.68 0.000 -.1041603 -.0810159
----------------+----------------------------------------------------------------
4.period |
urban|
Rural | -.0545424 .0181749 -3.00 0.003 -.0901657 -.0189191
Urban | -.1336001 .0051976 -25.70 0.000 -.1437876 -.1234126
----------------+----------------------------------------------------------------
5.period |
urban |
Rural | -.1943786 .0164296 -11.83 0.000 -.226581 -.1621761
Urban| -.243518 .0056055 -43.44 0.000 -.2545049 -.2325311
---------------------------------------------------------------------------------
