POPM 3240 Lecture Notes - Lecture 25: Logistic Regression, Linear Regression, Dependent And Independent Variables

• School: University of Guelph
• Department: Population Medicine
• Course: POPM 3240
• Professor: Andrea Jones- Bitton

Document Summary

Predict dependent variable y from explanatory variable x using equation. Determine which explanatory variables are important predictors of dependent variable and to what extent do they affect y. Must control for other explanatory variables (confounding) Once they are included in the model they are controlled. E. g. body weight, milk production, blood glc, wbc count. E. g. pregnant or not, hyperproteinemia or not, low birth weight or not. It is the outcome not the explanatory variable. Explanatory variable can be continuous, dichotomous, or categorical with either model. E. g. breed, nationality, education level, age (if categorized) All explanatory variables in the final model should be listed, and. E. g. smoking (no or yes), type of animal (0 for dairy, 1 for beef) Each explanatory variable should have regression coefficient (beta 1) and its standard error. Certain statistical assumptions need to be met and the fit of the model should be assessed using standard, accepted techniques.