Coefplot- Different outcomes, one independent variable over time

[Neg Kha]

Hi,

I have the following regressions:

  
 eststo clear
    

eststo sem1g: ivregress 2sls grade (housing=instrument) if semester==1, vce(r) //IV
eststo sem1t5: ivregress 2sls top5 (housing=instrument) if semester==1 , vce(r) //IV
eststo sem1t10: ivregress 2sls top10 (housing=instrument) if semester==1 , vce(r) //IV

eststo sem2g: ivregress 2sls grade(housing=instrument) if semester==2, vce(r) //IV
eststo sem2t5: ivregress 2sls top5 (housing=instrument) if semester==2 , vce(r) //IV
eststo sem2t10: ivregress 2sls top10 (housing=instrument) if semester==2 , vce(r) //IV

How can I plot the coefplot for the variable "housing" for all of the outcomes(grade, top5, top10) , separated by the semester? I have looked through the forum but I have not found anything that can help with what I exactly want.
To be exact, I want it to look like the following plot:

[Neg Kha]

Anyone who can help?

[Gobinda Natak]

Too bad that you don't provide an example to replicate, but if you can't get -coefplot- to do this for you, I would use -help postfile- to save the coefficients you are interested in into a new file, and then just plot the contents of that file with a regular -help twoway_dot- graph.

[Neg Kha]

Too bad that you don't provide an example to replicate, but if you can't get -coefplot- to do this for you, I would use -help postfile- to save the coefficients you are interested in into a new file, and then just plot the contents of that file with a regular -help twoway_dot- graph.

Thanks, here is an example of the data to replicate:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float(id grade top5 top10 semester housing instrument)
 1   .4430962       .25       .25 1 0 0
 1 -.13358703         0         0 2 0 0
 2  .03894464         0         0 1 1 1
 2 -1.2169484         0         0 2 1 1
 3  -.5927218         0         0 1 1 1
 3 -.14161418       .28       .28 2 1 1
 4   .2938398       .25       .25 1 0 0
 4  1.3441702         1         1 2 0 0
 5    -.42742       .25       .25 1 1 1
 5  -.6184571  .3333333  .3333333 2 1 1
 6  1.2497058         1         1 1 1 1
 6   .7064008        .4        .6 2 1 1
 7 -.29753825       .25       .25 1 1 1
 7  -.4294281         0         0 2 1 1
 8  -.5626787         0         0 1 1 1
 8  -.5956856         0         0 2 1 1
 9   .4098808         0        .5 1 1 1
 9  -.1975955 .14705883 .14705883 2 1 1
10   1.196604        .6        .8 1 1 1
10  .25098285       .25       .25 2 1 1
11   .8947548        .5        .5 1 1 1
11  -.5858306       .25       .25 2 1 1
12   1.406335         1         1 1 1 0
12  .27450433        .4        .4 2 1 0
13  -.5134795         0         0 1 0 0
13  .11924653  .3333333  .3333333 2 0 0
14  1.1021477         0         1 1 1 1
14 -.12462802  .3793103  .3793103 2 1 1
15  -.8376226         0         0 1 1 1
15  .14714698  .3421053  .4473684 2 1 1
16   .8084729  .3333333  .6666667 1 1 1
16  .54073656  .3333333  .3333333 2 1 1
17   -1.47094         0         0 1 1 1
17  -.4551369 .16666667 .16666667 2 1 1
18   .9733542        .5        .5 1 1 1
18   .3769173         0        .4 2 1 1
19   .7044744         1         1 1 0 1
19  -.5995409         0         0 2 0 1
20  1.5721848         1         1 1 1 1
20   .9947854  .8333333         1 2 1 1
end

[Gobinda Natak]

Thanks, if you do

coefplot sem*, keep(housing)

you'll need to label the y-axis somehow by hand, which isn't ideal, I guess, but as I said above, postfile and twoway dot should be able to do the trick