Same deal as Chapter 1 — real scenarios where you decide the approach. Now with the MLR twists: holding predictors constant, adjusted R², and the traps that come with more than one predictor.
Build the first model arrow_forward
Case Files walk you through framing, planning, crunching and deciding a two-predictor model. Coach the Rookie hands you a flawed MLR analysis to spot and fix — multicollinearity, dodgy interpretation, and more.
Win % from tackles and inside-50s. Frame it, choose the matrix approach, judge it, read it.
Start → LiveCoach the RookieA home-run coefficient went negative. Spot the multicollinearity and fix it.
Start → LiveCoach the RookieA write-up ranks predictors by coefficient size and forgets "holding constant". Fix both.
Start → LiveCoach the RookieR² rose when a junk predictor was added. Spot why that proves nothing.
Start → LiveCoach the RookieThe model's significant — but is every predictor? Read the table.
Start → LiveCoach the RookieA 956-run super-team projection. Spot the multi-predictor extrapolation.
Start →More MLR cases — interaction terms, model comparison — can slot straight in.