Example Using McNemar's Test 



  To study the association between retirement and heart disease, 127
victims of cardiac arrest were matched with 127 healthy individuals;
retirement status was then determined for each subject. Note that we
have 254 (2 * 127) subjects in the study.
             
		         Cardiac Arrest
			  Yes  |   No
              Retired
	      ---------------------------
		Yes       47       39
		No        80       88

  Note that each column adds up to 127, since there are 127 cardiac
 arrest victims and 127 who are not.

  Is there an association between retirement status and cardiac arrest?

  Suppose the raw data is available to us, so we can classify it in the
following manner:

                             Cardiac Arrest
			     ---------------
			   Retired  | Not Retired
 No Cardiac Arrest      
  ------------------------------------------------ 
    Retired                  27         12 
    Not Retired              20         68

 This tells us that, of the 47 subjects who were retired and suffered
cardiac arrest, 27 were matched with retired people who didn't suffer
cardiac arrest and 20 were matched with people who were not retired and
didn't suffer cardiac arrest. 

  Each entry in this table corresponds to a combination of response from
each pair, so the total of the numbers in this table is 127.

  NOTE: Given the information in the second table, we can construct the
first table (the row and column values in the second table become the
cell entries in the first table, etc).

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

> cardiac
            arr.ret arr.noret            # arr.ret = Cardiac Arrest
  noarr.ret      27        12            #           and Retired 
noarr.noret      20        68         

> mcnemar.test(cardiac)

         McNemar's chi-square test with continuity correction 

data:  cardiac 
McNemar's chi-square = 1.5312, df = 1, p-value = 0.2159