Combining binary variables into 0,0; 0,1; 1,0; 1,1 sets

[ksmielitz]

I conducted a chi-sq for my variables and got the following results:

Frequency Percent Row Pct Col Pct

Table of lowSD by PBClow lowSD PBClow 0 1 Total 0 32 22.70 66.67 30.48 16 11.35 33.33 44.44 48 34.04     1 73 51.77 78.49 69.52 20 14.18 21.51 55.56 93 65.96     Total 105 74.47 36 25.53 141 100.00

Which is great. Now I need to write the code to create NEW variables to represent the (0,0), (0,1), (1,0), and (1,1)...so that when I run a proc freq I would get something resembling:

(0,0) variable 1: 32

                      0: 109

(0,1) variable 1: 16

                      0: 125

(1,0) variable 1: 73

                      0: 68

(1,1) variable 1: 20

                      0: 121

My current code is this: (and I get all of my values in the 0 on the proc freq (see below the code)

*Two by two combo of SD and PBC;
if PBClow=1 and SDlow=1 then LowLow = 1; else lowlow =0;
if PBChigh=1 and SDhigh=1 then HighHigh = 1; else highhigh =0;
if PBClow =1 and SDhigh=1 then Compare = 1; else compare=0;
if PBChigh =1 and SDlow=1 then HighLow = 1; else highlow=0;

SAS Output

The FREQ Procedure

LowLow	Frequency	Percent	Cumulative Frequency	Cumulative Percent
0	141	100.00	141	100.00

HighHigh	Frequency	Percent	Cumulative Frequency	Cumulative Percent
0	141	100.00	141	100.00

Compare	Frequency	Percent	Cumulative Frequency	Cumulative Percent
0	141	100.00	141	100.00

HighLow	Frequency	Percent	Cumulative Frequency	Cumulative Percent
0	141	100.00	141	100.00

Thanks in advance for any help you can provide!

Kate

[PaigeMiller]

---

@ksmielitz wrote:

I conducted a chi-sq for my variables and got the following results:

Frequency Percent Row Pct Col Pct

Table of lowSD by PBClow lowSD PBClow 0 1 Total 0 32 22.70 66.67 30.48 16 11.35 33.33 44.44 48 34.04     1 73 51.77 78.49 69.52 20 14.18 21.51 55.56 93 65.96     Total 105 74.47 36 25.53 141 100.00

Which is great. Now I need to write the code to create NEW variables to represent the (0,0), (0,1), (1,0), and (1,1)...so that when I run a proc freq I would get something resembling:

(0,0) variable 1: 32

                      0: 109

(0,1) variable 1: 16

                      0: 125

(1,0) variable 1: 73

                      0: 68

(1,1) variable 1: 20

                      0: 121

I think you have a better choice than writing the code that comes next and then trying to debug it. If you create an output data set from PROC FREQ, you have the table you want, described just above, with just a little less difficult programming. Let PROC FREQ do most of the work for you.

[Reeza]

Look at the LIST option on the TABLES statement.

Table x * y / list out=want;