topic Re: How to make age a categorical variable? in SAS Programming

How to make age a categorical variable?

toesockshoe — Thu, 19 Jul 2018 20:43:51 GMT

I have a dataset with age ... values being like 15,16,17,...90 (all whole numbers), but sas makes this continuous (numerical). Is there a way for me to make this categorical?

Re: How to make age a categorical variable?

mkeintz — Thu, 19 Jul 2018 20:48:53 GMT

natural numbers are not continuous.

But I digress. What categories do you want?

Re: How to make age a categorical variable?

toesockshoe — Thu, 19 Jul 2018 20:50:58 GMT

so id want 21 to be 21 and 22 to be 22 and so on, but I just don't want them to be treated as integers. I'd like for them to be treated as different/independent groups.

Re: How to make age a categorical variable?

ballardw — Thu, 19 Jul 2018 20:54:25 GMT

Here's one example format:

Proc format;
value age10yr
0 - 9 =' 0-9'
10-19 ='10-19'
20-29 ='20-29'
30-39 ='30-39'
40-49 ='40-49'
50-59 ='50-59'
60-69 ='60-69'
70-79 ='70-79'
80-89 ='80-89'
90-99 ='90-99'
100-high ='100+'
;
run;

Use the format associated with the variable in Proc Print, Freq or what have you such as:

Format age age10yr. ;

The name of a format cannot end in a number to avoid headaches when using the notation to specify print width.

Re: How to make age a categorical variable?

mkeintz — Thu, 19 Jul 2018 20:58:51 GMT

By "treated as categories", I presume you mean the equivalent of a CLASS variable in a GLM, or other statistical procedures. If so, you don't have to modify the age variable. Instead just tell your analysis/tabulation procedure to treat age as a class var. In the GLM proc the statement would be

CLASS age;

which you would associate with the appropriate MODEL statement.

Re: How to make age a categorical variable?

PaigeMiller — Thu, 19 Jul 2018 21:49:04 GMT

@mkeintz wrote:

By "treated as categories", I presume you mean the equivalent of a CLASS variable in a GLM, or other statistical procedures. If so, you don't have to modify the age variable. Instead just tell your analysis/tabulation procedure to treat age as a class var. In the GLM proc the statement would be

CLASS age;

which you would associate with the appropriate MODEL statement.

And if we're talking about AGE, then yes the CLASS statement would cause AGE to be treated as a categorical variable, and @toesockshoe, most likely this is a poor way to analyze this variable.

Re: How to make age a categorical variable?

Astounding — Thu, 19 Jul 2018 22:51:08 GMT

As has already been mentioned, the CLASS statement does this in most cases.

If you find a case where the CLASS statement doesn't work, just create your own categorical variable based on AGE:

agecat = put(age, z3.);

Then use AGECAT instead of AGE in your program. Note that AGECAT will be character (not numeric).

Re: How to make age a categorical variable?

Reeza — Thu, 19 Jul 2018 23:13:41 GMT

Variables are numerical or character. A categorical variable is mostly defined by usage, but can typically be of either group. So why do you think you need a categorical variable?

If this is for a regression using GLM/LOGISTIC or that form you need to place the variable in a CLASS statement or create dummy variables manually. It won't matter if it's character or numerical when creating the dummy variables.

@toesockshoe wrote:

so id want 21 to be 21 and 22 to be 22 and so on, but I just don't want them to be treated as integers. I'd like for them to be treated as different/independent groups.