How to ignore negative numbers used in place of missing data when coding a binary dummy variable

[Calyn]

Hello, I am very new to SAS (just started learning it earlier this semester) and also relatively new to quantitative research (first year PhD student).

I am trying to run a regression analysis from a very large secondary dataset that is part of multiple waves (but I am only using one of the waves), that has replaced missing values with various negative numbers (-9 for not in wave, -3 for missing answer, -1 for refuse to answer, etc...). I am trying to create a binary variable from a multiple choice survey variable (because I'm not concerned with the degrees to which the behavior is present, only whether it is present or whether it is not present- so if it's present at al it will be 1, and if it's not present at all it will be 0), but I don't know how to handle the negative numbers so that the analysis just ignores them and doesn't include them for that variable.

I apologize if I'm not describing what I'm trying to do properly and if this is something that should be obvious on how to handle it. I appreciate any help figuring out the syntax.

[Astounding · Accepted solution]

The best approach would be to fix the data.  Get rid of the negative numbers, by changing them to missing values:

if var1=-9 then var1=.A;
else if var1=-3 then var1=.B;
else if var1=-1 then var1=.C;

As you can see, there are a variety of ways to code missing values.  You can assign missing values, yet still distinguish the reason why the variable is missing.

Regardless of what should happen next, that's going to be the right beginning.  And statistical procedures will recognize those values as missing and (where appropriate) discard the observation.

For more information, check the documentation for special missing values:

https://documentation.sas.com/?cdcId=pgmsascdc&cdcVersion=9.4_3.5&docsetId=lrcon&docsetTarget=p1xr9f...

[Calyn]

Thank you, I appreciate the help!