I have a patient group with two different categories of clinical features (say clinical group 1 and clinical group 2). And there are two diagnostic assays to analyze this patient group (say assay 1 and assay 2). According to results, there is a clear difference in giving positivity by each assay.
Assay 1 works better for both group 1 & 2. Assay 2 works better for group 1 than group 2. I want to show this difference by giving a p value.
Even though there are cases positive only by assay 1, there are very few (eg: only 1 from 11 cases) cases positive only by assay 2. Because majority of those are also responsive to assay 1 (it is expected within the study).
Therefore, those two groups are not mutually exclusive. Samples are overlapping. But those are independent test results.
[Note: Assay 1 & 2 are two different tests which I want to compare their response pattern among two different clinical groups (clinical group 1 & 2) mentioned].
Can crosstabulation (2X2) apply for this? Or is there any other statistical method to analyze this type of correlation?