Hi All,
I have a dataset containing a binary outcome(Pot=0/1) measured repeatedly for each subject (ID) (n=148) who are randomly assigned to a treatment or control group. The dependent variable is also a binary(Y=0/1) which is based a self report question subjects are asked at the end of the experiment.
-In the experiment there are 15 time periods for the treatment group, and 10 time periods for the control. subjects chose between the binary outcome variable each time period.
-The time periods can be grouped into three rounds of five sequential time periods (pre-treatment, treatment, post-treatment). The pre-treatment and post-treatment rounds are equivalent for treatment and control group.
-I expect to find that those assigned to the treatment group have a higher probability of choosing option 0 while those in the control group have a higher probability of choosing option 1.
-Further, I expect to find that those assigned the treatment have a higher probability of choosing option 0 in the self report question, while the control group participants have a higher probability of choosing option 1 in the self report.
-I was wondering if it possible to use xtgee Y .... , Family(binomial) link(logit) to examine the dynamics of choice over time for each subject?
Also is there any way to look at the change induced by the treatment effect over time in the last five time periods (The post-treatment rounds)?
- This is my first post on this forum so apologies if it is written unclearly.
- I am also quite new to Stata and econometrics so I may be completely incorrect by trying to use xtgee for this problem and in methodology in general.
Thank you in advance,
Best,
Sean
I have a dataset containing a binary outcome(Pot=0/1) measured repeatedly for each subject (ID) (n=148) who are randomly assigned to a treatment or control group. The dependent variable is also a binary(Y=0/1) which is based a self report question subjects are asked at the end of the experiment.
-In the experiment there are 15 time periods for the treatment group, and 10 time periods for the control. subjects chose between the binary outcome variable each time period.
-The time periods can be grouped into three rounds of five sequential time periods (pre-treatment, treatment, post-treatment). The pre-treatment and post-treatment rounds are equivalent for treatment and control group.
-I expect to find that those assigned to the treatment group have a higher probability of choosing option 0 while those in the control group have a higher probability of choosing option 1.
-Further, I expect to find that those assigned the treatment have a higher probability of choosing option 0 in the self report question, while the control group participants have a higher probability of choosing option 1 in the self report.
-I was wondering if it possible to use xtgee Y .... , Family(binomial) link(logit) to examine the dynamics of choice over time for each subject?
Also is there any way to look at the change induced by the treatment effect over time in the last five time periods (The post-treatment rounds)?
- This is my first post on this forum so apologies if it is written unclearly.
- I am also quite new to Stata and econometrics so I may be completely incorrect by trying to use xtgee for this problem and in methodology in general.
Thank you in advance,
Best,
Sean
