Dummy variables areg

[Renato Felix]

Hello all together,
I am new here so please let me know if I asked my question properly. Unfornately, I cannot post output due to confidentiality issues but I hope you can still help me. I am a beginner in Stata and I have a problem with dummy variables in an areg-regression:

I have a list of 33 dummy variables guts1-guts33 which stand for some variables of my dataset. I created them by "gen gutsd". Now I want to set some of them to 0 in my regression. I have tried to put the dummy variables in a macro:

global gutsch guts2-guts4
areg q2 $treatments date $controls if count==1 & $gutsch==0 , robust absorb(zip)

Problem: He seems to only set the first (guts2) to zero but i want him to set guts2, guts3 and guts4 to 0. If i put them in seperately it does the job but is not really an elegant way of doing it:

[CODE

• areg q2 $treatments date $controls if count==1 & guts2=0 & guts3 = 0 & guts4 =0, robust absorb(zip)

[/CODE]

I have not found anything but probably just did not look for the right thing.

Thanks in advance and best regards

Renato

[Phil Bromiley]

set trace on often helps in debugging such problems. My guess is that when Stata (an it normally) interprets $gutsch it ends up with a statement that says
& guts2-guts4==0 which doesn't work. If you're just starting with Stata, you should not worry much about elegance. A pile of conditional statements is a lot faster and easier than finding a cuter way.

You could perhaps do

egen temp=rowtotal(guts1-guts4)

And then just have the conditional temp==0 in your areg statement. This of course assumes guts take on the values 0 and 1.

You would need == in each of the equals in your second set of code. [This is part of the reason we like actual code - if you tried to run what you posted, it would not run so I assume you've run something else.]