topic Re: Table Count (N) and Percent (%)? in Statistical Procedures

Table Count (N) and Percent (%)?

ybz12003 — Wed, 03 Aug 2022 14:04:59 GMT

Hello,

I could like to create a table for listing the Counts and Percentages (One digit behind decimal) of three categories (Gender/Race/Premature) in two classes of Patients (ER and Hospital). Gender contains two subgroups, M and F. Race contains five subgroups, and I only need 1-3 subgroups. Premature contains Y/N subgroups. All the variables are numeric. The final result I would like to get is similar to the table shown below. Please advise how to approach it; Proc Freq, Proc tabulate, or something else. Thanks.

	Hospital Patient	ER Patient
Gender Male Female
Race White Black Hispanic
Premature (Y)

Re: Table Count (N) and Percent (%)?

PaigeMiller — Wed, 03 Aug 2022 14:06:37 GMT

I believe that the %TABLEN macro will do what you want.

https://communities.sas.com/t5/SAS-Communities-Library/Demographic-Table-and-Subgroup-Summary-Macro-TABLEN/ta-p/634030

Re: Table Count (N) and Percent (%)?

ybz12003 — Wed, 03 Aug 2022 14:25:51 GMT

The dis_order comment seems complicated. Does the data need to be sorted out first, then I have to figure out which needs to be listed in the order? Any other simpler procedure could be used?

Re: Table Count (N) and Percent (%)?

PaigeMiller — Wed, 03 Aug 2022 14:28:19 GMT

I still think the easiest thing to do is to use the %TABLEN macro. Other approaches would seem to require even more work.

Re: Table Count (N) and Percent (%)?

ballardw — Wed, 03 Aug 2022 14:28:56 GMT

You need to define exactly what "Race contains five subgroups, and I only need 1-3 subgroups." means. Does this mean to exclude any observations that have race not in the desired subgroups entirely from the table? Combine some of the groups (which you did not define) into those shown, or just ignore 2 classes and not include them in the report but count the observations in the Gender and Premature? This requirement may impose a bit of data restructuring depending on your answer

By "Premature contains Y/N subgroups. All the variables are numeric" does this mean that Premature has a numeric value of 1 for Yes and 0 for No? If not what is the actual coding?

If it is coded 1/0 then the SUM statistic gives you the number of Yes and Mean will give you a decimal percentage of yes

Re: Table Count (N) and Percent (%)?

ybz12003 — Wed, 03 Aug 2022 14:33:11 GMT

Yes.
Race has 1-5 subgroups, and I only need 1-3 groups. Group 4 and 5 will be discharged.
Premature = 1 is the one I need.

Re: Table Count (N) and Percent (%)?

ballardw — Wed, 03 Aug 2022 15:29:31 GMT

@ybz12003 wrote:
Yes.
Race has 1-5 subgroups, and I only need 1-3 groups. Group 4 and 5 will be discharged.
Premature = 1 is the one I need.

Does not quite answer the questions. Are the observations for race where it is 4 or 5 removed from the data entirely?

And what other value does Premature take? If it is not 0 we need to know.

Best, as has been requested for many projects is to include an actual example of the data set as data step code so we do not have to guess.

Then show the desired result given the example data set.

@ybz12003 wrote:
Yes.
Race has 1-5 subgroups, and I only need 1-3 groups. Group 4 and 5 will be discharged.
Premature = 1 is the one I need.

Re: Table Count (N) and Percent (%)?

ybz12003 — Wed, 03 Aug 2022 15:34:00 GMT

Race 4 = other, Race 5 = unknown
Premature = 1 or 0

Re: Table Count (N) and Percent (%)?

ybz12003 — Wed, 03 Aug 2022 15:36:02 GMT

I have the other extra ten variables that need to be counted. I would like to avoid figuring out the order of each category.

Re: Table Count (N) and Percent (%)?

Reeza — Wed, 03 Aug 2022 15:58:17 GMT

You don't need to figure out the order of each category that I can see unless you want it different than the default output. Do you want it different than the default? If so, how?