Linear Regression

Question 1

describe advantages of a linear model

Answer 1

• correct for multiple predictors simultaneously
  
  • in observational studies, usually the variables of interest are associated with variables that are not of interest
  • regression models allow the researcher to correct for these
• make predictions about the outcome
  
  • often researchers want to predict what will happen if one thing changes

Question 2

describe interpreting multiple regressiton models and detecting confounding

Answer 2

• if the slope estimate for a predictor changes after addition of another predictor, the association was confounded
• if the association gets weaker (closer to 0) there was positive confounding
  
  • temperature positively confounded the association between ice cream and sunburn
• if the negative association gets stronger (further from 0) there was negative confounding
  
  • UV index negatively confounded the association between temperature and sunburn