EPIDEMIOLOGY 530 Selection Bias & Information bias KEY
Subscribe to view the full document.
INTRODUCTION Results and interpretation of any epidemiologic study are subject to two types of error. Random error results from fluctuations around a true value because of sampling variability (chance) Systematic error (bias) produces differences between true value and observed value due to imperfections of study design, study implementation, or analysis of the data. There are three broad categories of bias Selection bias is defined as distortion in the effect estimate resulting from the manner in which subjects are selected for the study population, and/or selective losses from the study population prior to data analysis. Information bias is defined as a distortion in the effect estimate that occurs when measurement of either the exposure condition or the disease condition is systematically inaccurate. Confounding bias is defined as a situation when the estimate of the effect of the exposure of interest is distorted because it is mixed with the effect of an extraneous factor This exercise deals with the first two categories of bias: selection bias and information bias (misclassification bias).