This project applies multiple linear regression to predict the number of wins for professional baseball teams using a dataset of approximately 2,276 records spanning the years 1871 to 2006. Each observation represents a team's seasonal performance, adjusted to a 162-game schedule. The analysis follows a four-stage workflow: data exploration, data preparation, model building, and model selection.

Project 1

A practical walkthrough of Omitted Variable Bias (OVB) using the bwght dataset (n = 1,388). Covers the two conditions for OVB, direction of bias using the 2×2 matrix, and a side-by-side regression comparison via stargazer — showing how omitting family income overstates the negative effect of smoking on infant birth weight. Built with R and the wooldridge package.

Code Along 9