Hope you are enjoying the coziness of the weekend or early Monday already!.
I have a data set with two by two-way design. I got the overall difference-in-difference effect from regression, but I am not certain what is the best way to write a syntax to get multiple DID effects of each subgroup.
The model is about the change of Grade Point Average in Time 0 and Time 1 between controlled and treated groups. I am more interested, however, in the changes of GPA in each race/ethnicity & gender (i.e, female African, male African, female Hispanic, male Hispanic, female White, male White, and more).
The interpretation of coefficients is not a big deal when I have a dummy variable of gender. However, I am getting lost as the model has many dummy variables of gender and race/ethnicity in addition to the DID effect. Should it be the best way to define subgroups and run regression multiple times to get DID effects of a subgroup? or is there a better idea I can borrow?
regards,
Model I used to have Overall DID effect:
reg GRADE_POINT i.T##i.TREATED [pw = CREDIT_HR]
Model i am considering to use for DID effect of each subgroup:
reg GRADE_POINT i.T##i.TREATED [pw = CREDIT_HR] if GENDER == [value] & RE == [value]
----------------------- copy starting from the next line -----------------------
------------------ copy up to and including the previous line ------------------
I have a data set with two by two-way design. I got the overall difference-in-difference effect from regression, but I am not certain what is the best way to write a syntax to get multiple DID effects of each subgroup.
The model is about the change of Grade Point Average in Time 0 and Time 1 between controlled and treated groups. I am more interested, however, in the changes of GPA in each race/ethnicity & gender (i.e, female African, male African, female Hispanic, male Hispanic, female White, male White, and more).
The interpretation of coefficients is not a big deal when I have a dummy variable of gender. However, I am getting lost as the model has many dummy variables of gender and race/ethnicity in addition to the DID effect. Should it be the best way to define subgroups and run regression multiple times to get DID effects of a subgroup? or is there a better idea I can borrow?
regards,
Model I used to have Overall DID effect:
reg GRADE_POINT i.T##i.TREATED [pw = CREDIT_HR]
Model i am considering to use for DID effect of each subgroup:
reg GRADE_POINT i.T##i.TREATED [pw = CREDIT_HR] if GENDER == [value] & RE == [value]
----------------------- copy starting from the next line -----------------------
Code:
* Example generated by -dataex-. For more info, type help dataex clear input double PIDM float(T TREATED) byte(GENDER RE) str5 CRN float GRADE_POINT double CREDIT_HR 604323 0 0 1 7 "40002" 0 3 611455 0 0 1 7 "40002" 2.7 3 612230 0 0 1 7 "40003" 3.7 3 607659 0 0 0 5 "40003" 2 3 603693 0 0 0 7 "40003" 2.7 3 611084 0 0 1 7 "40003" 2 3 604251 0 0 0 7 "40003" 0 3 609437 0 0 1 5 "40004" 2 3 604487 0 0 1 7 "40004" 0 3 604836 0 0 1 7 "40004" 2 3 603273 0 0 1 7 "40004" 3.7 3 606161 0 0 0 5 "40006" 2.7 3 615142 0 0 0 5 "40006" 2 3 606312 0 0 1 7 "40006" 1.7 3 606242 0 0 0 7 "40007" 2.7 3 612811 0 0 0 5 "40007" 1.7 3 605898 0 0 1 5 "40009" 2.7 3 608254 0 0 0 7 "40010" 4 3 610725 0 0 1 7 "40011" 2 3 603484 0 0 0 5 "40012" 0 3 603611 0 0 0 2 "40013" 3.3 3 606323 0 0 0 7 "40013" 3.7 3 611313 0 0 0 7 "40013" 1.7 3 606587 0 0 0 7 "40014" 4 3 607473 0 0 1 7 "40014" 3.7 3 613507 0 0 0 7 "40016" 4 3 604475 0 0 0 7 "40017" 3.7 3 606806 0 0 0 7 "40018" 3 3 610340 0 0 0 7 "40019" 4 3 616039 0 0 0 5 "40019" 3 3 end label values T T label def T 0 "T0", modify label values TREATED TREATED label def TREATED 0 "Controlled", modify label values GENDER GENDER label def GENDER 0 "Female", modify label def GENDER 1 "Male", modify label values RE RE label def RE 2 "Hispanic/Latino", modify label def RE 5 "Black or African American", modify label def RE 7 "White", modify
Comment