topic Re: Problems with If-Then Statements in Statistical Procedures

Problems with If-Then Statements

bnd — Sat, 22 Feb 2020 05:56:03 GMT

Hello,

I am having some issues with the if-then statement. I am trying to make white and black into a new variable called white_black so that I can compare the two using just one variable. Both the white and black variables are coded as 1 if a person is that race and 0 if they are not. There are a total of 320 white individuals and 76 black individuals. When I place the variables into the if-then statement, it seems that there are 2 fewer people included in either group. For example, there may be 320 for white and 74 for black. I am not sure why this is happening. My code is provided below.

if white = 1 then black_white = 1

if black = 1 then black_white = 0

Re: Problems with If-Then Statements

ed_sas_member — Sat, 22 Feb 2020 06:03:10 GMT

Hi @bnd
Could you please run a proc freq to look at the data and share the result?

Proc freq data=have;
Tables white * black / nocum nopercent;
Run;

Re: Problems with If-Then Statements

bnd — Sat, 22 Feb 2020 06:20:55 GMT

Hi @ed_sas_member,

I have included the results in the attachment below.

Re: Problems with If-Then Statements

ed_sas_member — Sat, 22 Feb 2020 08:12:39 GMT

Hi @bnd

Thank you for the output.

It seems that 2 people have white=1 and Af_amer=1 -> maybe you should clean these data first.

Re: Problems with If-Then Statements

Kurt_Bremser — Sat, 22 Feb 2020 08:27:33 GMT

Your "problem" is caused by 7 people who consider themselves neither af_amer nor white, and 2 who think they are af_amer and white.

So instead of 401, your if/then will add up to 396.

Correction: you'll have 7 missing, 320 "white", and 74 "black", because with the double identification white overrides black.

PS your question is a nice illustration of Maxim 3 and the usefulness of proc freq.

Re: Problems with If-Then Statements

bnd — Sat, 22 Feb 2020 17:57:59 GMT

Thanks! @Kurt_Bremser @ed_sas_member

I kept trying to figure out what was going on exactly. After running the frequency table and reading over both of your comments, everything made perfect sense.

Thanks so much for your help!