/* SAS code to analyze an r by c contingency table */
/* and do chi-squared test of independence         */

/* We look at the snoring / heart disease data set: */

DATA snoreheart;
INPUT heart $ snoring $ number;
cards;
yes never 24
yes occas 35
yes almostevery 51
no never 1355
no occas 603
no almostevery 416
;
run;

/* We test for independence of the two classifications using */
/* the chisq option in PROC FREQ.                            */

/* In the TABLES statement below, the first variable listed is the row variable */
/* and the second variable listed is the column variable.                       */

PROC FREQ DATA = snoreheart ORDER=DATA;
TABLES heart*snoring / chisq EXPECTED;
WEIGHT number;
run;

/* The expected counts are given in each cell below the observed counts. */

/* P-value given as less than .0001. */

/**************************************************************/

/* SAS code to analyze an r by c contingency table */
/* and do chi-squared test of homogeneity          */

/* We look at the voter survey data set: */

DATA voter;
INPUT sample $ category $ number;
cards;
Upstate Approve 61
Upstate Neutral 50
Upstate Disapprove 39
Midlands Approve 56
Midlands Neutral 40
Midlands Disapprove 54
Coastal Approve 47
Coastal Neutral 37
Coastal Disapprove 66
;
run;

/* We test for homogeneity using  */
/* the chisq option in PROC FREQ. */

/* In the TABLES statement below, the first variable listed is the row variable */
/* and the second variable listed is the column variable.                       */

PROC FREQ DATA = voter ORDER=DATA;
TABLES sample*category / chisq EXPECTED;
WEIGHT number;
run;

/* The expected counts are given in each cell below the observed counts. */

/* The P-value is 0.027. */

/**************************************************************/

/* SAS code to do Fisher's exact test */
/* on a made-up small data set        */
/* (3 by 3 table)                     */

DATA small;
INPUT sample $ category $ number;
cards;
Upstate Approve 6
Upstate Neutral 5
Upstate Disapprove 3
Midlands Approve 5
Midlands Neutral 4
Midlands Disapprove 5
Coastal Approve 4
Coastal Neutral 3
Coastal Disapprove 6
;
run;

/* We use PROC FREQ with the FISHER option. */

/* In the TABLES statement below, the first variable listed is the row variable */
/* and the second variable listed is the column variable.                       */

PROC FREQ DATA = small ORDER=DATA;
TABLES sample*category / FISHER EXPECTED;
WEIGHT number;
run;