AI summary by TwinMind
Multiple Regression in R, Step-by-Step!!!
StatQuest with Josh Starmer
7m 42s
Introduction and Setup
- StatQuest tutorial by Josh Stormer from UNC Chapel Hill Genetics Department
- Comparing simple vs multiple regression in R
- R code available on StatQuest website (statquest.org)
- Dataset: measurements for size, weight, and tail variables
Simple Regression Analysis
- Objective: Use weight to predict size
- Step 1: Plot data first
- Weight on x-axis, size on y-axis
- Visual inspection shows linear relationship exists
- Confirms regression analysis is appropriate
- Step 2: Use lm() function (linear model)
- Syntax: size ~ weight (tilde character)
- R automatically adds y-intercept and slope terms
- Uses least-squares to minimize squared residuals
- Results Summary:
- R-squared = 0.613
- P-value = 0.012
- Weight does "pretty good job" predicting size
- For simple regression, use multiple R-squared (not adjusted)
- Visualization: Add fitted line using abline() function
Multiple Regression Analysis
- Objective: Use weight AND tail to predict size
- Step 1: Plot all data relationships
- R plots all columns against each other automatically
- Size vs weight, size vs tail, weight vs tail relationships shown
- Both weight and tail correlated with size (good predictors)
- Weight and tail correlated with each other (provide similar information)
- Suggests might only need one predictor, not both
- Step 2: Use lm() function for multiple variables
- Syntax: size ~ weight + tail (plus symbol for multiple predictors)
- R adds y-intercept, slope1, and slope2 terms
- Fits plane to data instead of line
Multiple Regression Results Analysis
- Key Metrics: R-squared, adjusted R-squared, p-value all look good
- Important: For multiple regression, focus on adjusted R-squared
- Model Comparisons:
- Multiple regression (weight + tail) vs simple regression (tail only)
- P-value shows no significant improvement
- Using both isn't significantly better than tail alone
- Multiple regression (weight + tail) vs simple regression (weight only)
- P-value shows significant improvement
- Using both significantly better than weight alone
- Multiple regression (weight + tail) vs simple regression (tail only)
Final Conclusions
- Using weight + tail to predict size is good overall
- Could save time by only measuring tail length (skip weighing mice)
- Tail alone performs nearly as well as both variables combined
