topic Re: Using proc freq to categorize in SAS Programming

Using proc freq to categorize

ChuksManuel — Sat, 02 Nov 2019 22:50:41 GMT

Hello forum,

I want to classify variables based off their responses - 0 to 9 into "A" , "B" and "C".

Especially, i want people with the "0" response that is less than 30% in the "C" category and people with "0" response between "30% -80%" in the "B" category and people with >80% in the C category.

Please how do i go about this.

Any help or idea is appreciated.

proc freq data=two;
table T1	T2	T3	T4; run;

Re: Using proc freq to categorize

mkeintz — Sun, 03 Nov 2019 01:20:46 GMT

How do you intend to represent the variable classification> Will you produce three datasets (A,B, and C) with qualifying variables in each? Do you just want a report table with each variable listed along with the classification?

Alsol, do any of your original variables ever have missing values? If so, do the 30% and 80% cutpoints include counting the missing values, or not?

Re: Using proc freq to categorize

ChuksManuel — Sun, 03 Nov 2019 01:44:00 GMT

Thank you so much.

I hope them to be in 3 different datasets.

I am so confused on how to do that with proc freq.

Re: Using proc freq to categorize

Reeza — Sun, 03 Nov 2019 07:07:40 GMT

Are you classifying people/individual records or variables?

proc freq summarizes variables not individuals.

plans what does I want people with 0 response that is less than 30% in the C and people with the 0 response in the 30 to 8 as B.
how does one person have more than one response?

@ChuksManuel wrote:

Hello forum,

I want to classify variables based off their responses - 0 to 9 into "A" , "B" and "C".

Especially, i want people with the "0" response that is less than 30% in the "C" category and people with "0" response between "30% -80%" in the "B" category and people with >80% in the C category.

Please how do i go about this.

Any help or idea is appreciated.
proc freq data=two;
table T1	T2	T3	T4; run;

Re: Using proc freq to categorize

PaigeMiller — Sun, 03 Nov 2019 16:33:45 GMT

PROC FREQ produces only one data set. You could take the results of PROC FREQ and use them to help create other data sets, however, your stated rules are not clear and definitive enough to create code that winds up which creates these data sets. How can "0" less than 30% be in C, but other "0" between 30 and 80 be in "B"? How do we tell the "0" that are in C apart from "0" that wind up in "B". Explain this in much more detail, give us an example.