Now imagine that cases underestimated their exposure and, as a result, 20% of exposed cases were falsely classified as unexposed, but that the classification of controls was not affected. (a) Is this...

Now imagine that cases underestimated their exposure and, as a result, 20% of exposed cases were falsely classified as unexposed, but that the classification of controls was not affected.

(a) Is this type of misclassification random or systematic? Is it differential or non-differential? Why? (b) What effect would it have had on the results of the study? (c) Compare your answer to (b) with that in Table 7.7, where cases systematically overestimated their exposure.

Table 7.7