topic Re: Wilcoxon signed rank and Mc Nemar tests in Statistical Procedures

Wilcoxon signed rank and Mc Nemar tests

Liamb — Wed, 16 Aug 2023 15:38:19 GMT

Hello,

Is it possible to perform the Paired Wilcoxon test and the McNemar using the Paired variable in the dataset below? Or do I have to transform the data to use proc univariate (diff)?

Thanks

data match;
  input Id $  case_control $ match_caseid $ paired age gender $ weight ethnic $;
datalines;
A1 Case    A1 1 20 F 56 Balck
A2 Control A1 1 23 F 60 Black
A3 Control A1 1 22 F 65 Black
B1 Case    B1 2 30 M 70 Asian
B2 Control B1 2 35 M 47 Asian
B3 Control B1 2 33 M 75 Unknown
C1 Case    C1 3 40 M 80 White
C2 Control C1 3 38 M 73 White
D1 Case    D1 4 47 M 50 Other
D2 Control D1 4 41 M 58 Other 
D3 Control D1 4 50 M 49 Unknown
D4 Control D1 4 63 M 77 Other
run;

Re: Wilcoxon signed rank and Mc Nemar tests

ballardw — Wed, 16 Aug 2023 16:01:27 GMT

How do you expect to use a PAIRED test with THREE values?

Pair very strongly implies 2. And in the case of paired statistics pretty much a requirement.

Perhaps you need to rephrase exactly the meaning of the resulting test instead of requiring a specific test.

Re: Wilcoxon signed rank and Mc Nemar tests

Liamb — Wed, 16 Aug 2023 16:20:41 GMT

Case were match to 1, 2 or 3 controls so a case with 1 control corresponds
to 1 pair, a case with 2 controls corresponds to 2 pairs and a case with 3
controls corresponds to 3 pairs.

Re: Wilcoxon signed rank and Mc Nemar tests

Liamb — Thu, 17 Aug 2023 10:05:04 GMT

Cases were matched with 1 or several controls (match_caseid) and I want to perform Wilcoxon signed rank test and the McNemar using case_control and Paired variables in the dataset below? It is possible?

Thanks

data match;
  input Id $  case_control $ match_caseid $ paired age gender $ weight ethnic $ ;
datalines;
A1 Case    .  1 20 F 56 Balck
A2 Control A1 1 23 F 60 Black
A3 Control A1 1 22 F 65 Black
B1 Case    .  2 30 M 70 Asian
B2 Control B1 2 35 M 47 Asian
B3 Control B1 2 33 M 75 Unknown
C1 Case    .  3 40 M 80 White
C2 Control C1 3 38 M 73 White
D1 Case    .  4 47 M 50 Other
D2 Control D1 4 41 M 58 Other
D3 Control D1 4 50 F 49 Unknown
D4 Control D1 4 63 M 77 Other
run;

Re: Wilcoxon signed rank and Mc Nemar tests

Ksharp — Thu, 17 Aug 2023 11:50:18 GMT

I don't understand, there is not Paired Wilcoxon test, Wilcoxon signed rank is just the non-parameter version of pair ttest ,it is for continuous variable ,not category variable.

But McNemar test is for the paired category variable(a.k.a freq/count data).
There are two different statistical method.

And since it is about stat question,better post it at stat forum:

https://communities.sas.com/t5/Statistical-Procedures/bd-p/statistical_procedures

and @StatDave maybe give you a help.