X<-table(colds$age12,colds$age14)
summary(mcNemar(X))R v SAS McNemar’s test
McNemar’s test; R and SAS
In R, the mcNemar function from the epibasix package can be used to perform McNemar’s test.
The FREQ procedure can be used in SAS with the AGREE option to run the McNemar test, with OR, and RISKDIFF options stated for production of odds ratios and risk difference. These options were added as epibasix::mcNemar outputs the odds ratio and risk difference with confidence limits as default. In contrast to R, SAS outputs the Kappa coefficients with confident limits as default.
proc freq data=colds;
tables age12*age14 / agree or riskdiff;
run;When calculating the odds ratio and risk difference confidence limits, SAS is not treating the data as matched-pairs. There is advice on the SAS blog and SAS support page to amend this, which requires a lot of additional coding.
R is using Edward’s continuity correction with no option to remove this. In contrast, there is no option to include Edward’s continuity correction in SAS, but this can be manually coded to agree with R. However, its use is controversial due to being seen as overly conservative.
R’s use of the continuity correction is consistent with other functions within the epibasix package, which was categorised as ‘High Risk’ by the Risk Assessment Shiny App created by the R Validation Hub. Risk is quantified by the app through a number of metrics relating to maintenance and community usage. It was found that the author is no longer maintaining the package and there was no documentation available for certain methods used. Therefore, the use of the epibasix package is advised against and other packages may be more suitable.
The mcnemar.test function in the stats package provides the option to remove continuity corrections which results in a match with SAS. This function does not output any other coefficients for agreement/difference in proportions etc. but (if required) these can be achieved within other functions and/or packages.
mcnemar.test(X, correct = FALSE)