HSCI 330 Lecture 6: Module 6
Document Summary
Section 9-1 (skip box general formulation of misclassification bias p. 232) Section 9-2 (skip box sensitivity and specificity parameters are conditional prob. p. 239) and read up to page 245 and read section independent misclassification of both exposure and disease. Misclassifying disease status: misclassification of disease status may occur from any of the following sources: Answers: 2, 7, 6/8=0. 75 = 75, 7/10=0. 70 = 70% Misclassification of exposure status can occur because of imprecise measurement of exposure: misclassification of exposure status may occur from any of the following sources: Imprecise measurement: subject self-report reluctance of subjects to be truthful in reporting exposures relating to behaviours, ex. Studies that measure food intake, sexual behaviour, illegal drug use. Interviewer bias interviewer probes more thoroughly about exposure for case than control. Incorrect coding of exposure data: a misclassification table provides a convenient summary of how exposure can be misclassified from true exposure to observed exposure. Finally, exposure data can be coded incorrectly in a database.