Resources · Business Analytics
Multiple regression
What it means
Multiple regression estimates how an outcome is associated with several predictors while holding the other included predictors constant.
How to do it
1. Define the question
Name the outcome, predictors, population and period. Explain why each variable belongs in the model.
2. Check the data and model
Inspect missing values, unusual observations, functional form, residuals and predictor overlap before interpreting coefficients.
3. Interpret in context
Report direction, size, units and uncertainty. Keep the interpretation conditional on the other included predictors.
Common mistakes & limitations
Association is not causation
A fitted coefficient does not by itself prove a causal effect. Omitted variables, measurement choices and model assumptions can change the result.
Use this with my assignment
Use this in an assignment →