EPID 301 Chapter Notes - Chapter 6: Categorical Variable, Blood Pressure, Underweight
Document Summary
Chapter 6: measurement error that leads to misclassification. Unlike random error, systematic error does not diminish with larger sample size. Comes from defects in study design that remain unchanged no matter how large the sample. Leads to bias- estimates that systematically over- or underestimate the true values. There are two types of study defects that can introduce systematic error: measurement error and selection error. To calculate an epidemiological estimate, you need to assign values to 1 or more variables for all members of the sample that forms the basis of your estimate. Assigning the values to variables is not always a perfect process (possible that a person with a disease may be incorrectly classified as not having that disease and vice versa) Errors of classification-misclassification is particularly important type of measurement error in epidemiology. Different types of data require different ways of thinking about measurement. Nominal data: conform to categories that can be named (ex.