8.1 Apply the McNemar test to table 8.3 Iterpret (Using SAS. Give the output please)

Solution:

McNemar Test

Below is the code to create the data in SAS (1 stands for “Diabetes” , 0 stands for “No Diabetes”)

data Matched;
input MI_control MI_cases count;
datalines;
1 1 9
1 0 16
0 1 37
0 0 82
;
This is how the data looks:



After creating the data, we run the SAS code below to create the McNemar’s test:

proc freq; weight count;
tables MI_control*MI_cases / agree; exact mcnem;
run;

The SAS output is as below:

The FREQ Procedure
Frequency
Percent
Row Pct
Col Pct
Table of MI_control by MI_cases
MI_control
MI_cases
0 1 Total
0 82
56.94
37
25.69
119
82.64
68.91
83.67
31.09
80.43
1 16
11.11
64.00
16.33
9
6.25
36.00
19.57
25
17.36
Total 98
68.06
46
31.94
144
100.00
Statistics for Table of MI_control by MI_cases
McNemar's Test
Statistic (S) 8.3208
DF 1
Asymptotic Pr > S 0.0039
Exact Pr >= S 0.0055
Simple Kappa Coefficient
Kappa 0.0369
ASE 0.0789
95% Lower Conf Limit -0.1178
95% Upper Conf Limit 0.1915
Sample Size = 144


Since p-value of the McNemar’s test is less than 0.05 ( 5 % level of significance), we reject the
null and say that row marginal frequencies and column marginal frequencies are different.



Wald confidence...