Misclassification

1.  Misclassification arises from errors in measuring study data. 2.  Nondifferential misclassification results from random errors that occur roughly equally within a study population. 3.  Nondifferential misclassification usually yields measures of association that are closer to the null value of 1.0. 4.  Differential misclassification results from systematic errors that occur preferentially within a subset of a study population. 5.  Differential misclassification may obscure true associations, falsely amplify associations, or create false associations 6.  Uniform data collection procedures help to minimize the possibility for differential misclassification.