New to STATA: separating numeric value variable

[Prapti Mallhi]

Hello,

I am very new to STATA and am trying to learn by myself. However, I am stuck at this point:

How can I split a numeric variable 'marstat'. It's label is current marital status - It's components are never married, widowed, etc. and I want to generate a new variable that only consists of 'never married'. Any help or guidance is appreciated.

Thank you in advance.

[Andrew Musau]

Is it a string variable? If so

gen never_married= marstat== "never married"

If it is a numeric variable and e.g., never married takes on values 2

gen never_married= marstat== 2

[Jocelyn Cherry]

What is it coded as?
I am also a learner, but this should work:

egen varname= anymatch (marstat), value (3)
tab varname, miss

where varname is a name you choose for your new variable and 3 must be the numeric value never married is 'tagged' with (ie if never married is coded 1 in your data, you need to change 3 to 1).

[Prapti Mallhi]

Thank you so much for your prompt reply!

I tried both -
and the second code worked, however, it's giving me 2 values. I only need it to be the '1' with frequency of 169.814. any ideas on how I could do that?

Thank you for all your help!

[Andrew Musau]

EDITED:

I tried both -
and the second code worked, however, it's giving me 2 values. I only need it to be the '1' with frequency of 169.814. any ideas on how I could do that?

You have generated an indicator for never married (=1 if an individual in the sample is never married and zero otherwise). So if you want to run a regression with only the sample of never married individuals, you use the -if- qualifier

regress y x1 x2 ... xN if never_married

You do not need to drop the other observations. But if you must

drop if !never_married
tab never_married
tab marstat

[Prapti Mallhi]

Got it, thank you!