HESC 401 Chapter Notes - Chapter 15: Recall Bias, Confounding, Reporting Bias
Document Summary
Any error in design that results in a false association. There are many different types of bias but the three most prevalent in epidemiological studies are recall bias, selection bias, and information bias. Having selection bias may present findings that are not representative of an entire population. Occurs when the data acquisition about the subjects is inadequate or done improperly; there is then an error when explaining a causal relationship. Information bias gives way to misclassification bias: subjects are misclassified as controls (when they should be cases) or as cases (when they should be controls). Differential misclassification: rate of misclassification differs in different study groups: cases are misclassified more often than controls (or vice versa), leads to an apparent but non-existent association. Non-differential misclassification: misclassification as a result of errors in data collection methods: usually, this results in relative risk is closer to 1. 0, meaning detection of an existent association may be difficult.
