Understanding and checking the assumptions of linear regression: A primer for medical researchers

Abstract

Linear regression (LR) is a powerful statistical model when used correctly. Because the model is an approximation of the long-term sequence of any event, it requires assumptions to be made about the data it represents in order to remain appropriate. However, these assumptions are often misunderstood. We present the basic assumptions used in the LR model and offer a simple methodology for checking if they are satisfied prior to its use. In doing so, we aim to increase the effectiveness and appropriateness of LR in clinical research.