Misclassification in 2 X 2 Tables

摘要: Although in the derivation of chi-square test (and related techniques) classifications 2 X table are assumed to be correct, there many practical problems where mistakes classification going made. The important question then arises: What effects will misclassification have on conclusions drawn from usual significance tests? principal purpose this paper is answer question. In medical field some that involve almost no risk error, for example, categories "lived" and "died". On other hand, more complex diagnoses clinician realizes a considerable may vary great deal depending disease under study, existence availability diagnostic tests, factors. For diseases direction missing actual cases. misdiagnosis due which "mimic" characteristics study. Even when precise possible time or cost factors necessitate use cruder methods. large survey mental health, it rarely psychiatric all respondents. Instead necessary classify respondents as "Health Problems" "Not Health basis information schedule. One procedure such situations construct health scale (using subsample "problem" "non-problem" cases) thereafter individuals by whether they fall above below critical point scale. Such well lead appreciable misclassifications both directions.