prtest for paired proportions

[Inaamul Haq]

Hi
I want to test the difference between two paired proportions. I need the Confidence Interval for the difference between proportions. I have a paired design wherein each individual with known disease status d is administered two tests (t) with the test results in y. I need to test the hypothesis that the (proportion y = 1 with test 1) - (proportion y =1 with test 0) = 0 in people with d == 1. Part of the data is shown below.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float id byte(test d y)
 1 1 0 1
 1 2 0 0
 2 1 0 1
 2 2 0 1
 3 1 0 0
 3 2 0 0
 4 1 0 0
 4 2 0 0
 5 1 0 0
 5 2 0 1
 6 1 0 0
 6 2 0 0
 7 1 0 0
 7 2 0 0
 8 1 0 0
 8 2 0 1
 9 1 0 0
 9 2 0 1
10 1 0 0
10 2 0 0
end
label values d ynlbl
label def ynlbl 0 "no", modify
label values y negposlb
label def negposlb 0 "-", modify
label def negposlb 1 "+", modify

When I read the prtest help, I understand that the difference between paired sample proportions needs the cluster() and rho() to be specified. I realize the cluster() in my case is cluster(id) but I am not sure about the rho(). How do I calculate the rho() for my data?

Is it

corr y test if d == 1

Please provide guidance.

[Dirk Enzmann]

The description of your problem does not fit to the data you are showing: There is no case with d == 1 and the values of "test" are 1 and 2, not 1 and 0.

If I would assume that you want to compare the proportions of 1 in y across the measurements test == 1 and test == 2 if d == 0 (sic!), you could reshape your data from long to wide and use -prtest- for two variables (paired test):

reshape wide y, i(id) j(test)
prtest y1=y2 if d==0

[Inaamul Haq]

Thank you Dirk Enzmann . Sorry to have pasted the wrong data.
I am pasting the correct data here again.

* Example generated by -dataex-. To install: ssc install dataex
clear
input float id byte(d y) float test
 1 0 1 0
 1 0 0 1
 2 0 1 0
 2 0 1 1
 3 0 0 0
 3 0 0 1
 4 0 0 0
 4 0 0 1
 5 0 0 0
 5 0 1 1
 6 0 0 0
 6 0 0 1
 7 0 0 0
 7 0 0 1
 8 0 0 0
 8 0 1 1
 9 0 0 0
 9 0 1 1
10 0 0 0
10 0 0 1
end
label values d ynlbl
label def ynlbl 0 "no", modify
label values y negposlb
label def negposlb 0 "-", modify
label def negposlb 1 "+", modify

But I don't think reshaping the data to wide would solve my problem. I am not sure but

prtest y1 = y0 if d == 0

in wide data and

prtest y if d == 0, by(test)

in long data would essentially be the same.

[Dirk Enzmann]

You are correct, you need no -reshape-. Note that in your example d is still 0 for all cases.

[Inaamul Haq]

Well I realised mcc reports the MCNemar's test as well as the 95% CI for the difference in proportions

mcc casevar controlvar