I think this is a rather basic question, but I just realize that I don't quite understand the term continuity correction.

I use R and found the same syntax correct=TRUE for both chisq.test and mcnemar.test. Are they referring to different continuity correction method?

I have heard of the yate's continuity for pearson chi-square test is not very popular now as it may "over-adjust" the result, then how about that for McNemar's chi-square test?



Another reason that continuity corrections for contingency tables have gone out of fashion is that they only make a noticeable difference when the cell counts are small, and modern computing power has made it feasible to calculate 'exact' p-values for such tables.

Exact tests for 2×2 tables for R are provided by the exact2x2 package written by Michael Fay. There's an accompanying vignette about the exact McNemar's test and matching confidence intervals.

  • 1
    $\begingroup$ (+1) I discovered this package only recently. It looks interesting. $\endgroup$ – chl Jan 22 '11 at 11:17
  • 4
    $\begingroup$ His paper about it got into Biostatistics: dx.doi.org/10.1093/biostatistics/kxp050 $\endgroup$ – onestop Jan 22 '11 at 11:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.