Tabulate Multiple Longitudinal variables

[Dom_Cav]

Hi friends,

I currently have a dataset with variables like this:

Partner1 Partner2 Partner3 Job1 Job2 Job3

where both Partner and Job are categorical variables (Job=Yes/No, Partner = Partnered/Not Partnered), and the number after the variable represents the time point in which the variable was observed.

I want to make a cross tabulated table that shows Partner status by Job status at all points in time.

I want it to look like this:

where each cell has the count of people in each group. Ive been using proc tabulate, and that seems to work if the row variable is the same at all time points, but im not sure how to get it to consider a new variable for another time point.

Any help is appreciated.            

[Ksharp · Accepted solution]

You have to change the data structure before using proc tabulate.

data want;

 set have;

job=Job1;output;

job=Job2;output;

job=Job3;output;

run;

proc tabulate data=want;

class job Partner1 Partner2 Partner3;

table job,Partner1 Partner2 Partner3;

run;

[ballardw]

If your data has "time" variable, a variable to indicate partner status and job status similar to:

Time   Partner   Job

1          Yes         Yes

1          No           Yes

2          Yes          No

etc.

then

proc tabulate data=have;

    class time partner job;

    table job,

            time*Parter*n='';

run;

should work unless you have missing values somewhere