Module · Framing✅Knowledge checkCheck: can you read a fitted line?5 questions · Pass at 60%1. In a regression of sales on TV budget, what does the intercept represent to a marketer?Base sales — what the market sells with no TV budget at allThe smallest TV budget worth buyingThe error the model makes on an average marketThe share of sales that TV is responsible for2. R² of 0.61 tells you that the model…is right 61% of the timeexplains 61% more than predicting the average every timepredicts 61% of markets within one thousand unitswill be wrong by 61% on a new market3. You must give a marketing lead ONE number they will plan against. Which?R²RMSE, in unitsThe slopeThe sum of squared errors4. The residual plot shows a clear arc rather than a shapeless band. That means…the data has outliers that should be removedthe model is missing structure — it is wrong in a pattern, not at randomR² has been overstated and should be recomputedthe sample is too small to model5. Errors are far more spread out for high-budget markets than low-budget ones. Why does that matter commercially?It means the model cannot be used at allIt only affects R², not the forecastThe forecast is least reliable exactly where the most money is being spentIt means low-budget markets should be dropped from the dataCheck answers← Back to Regression I — predicting a numberNext topic →
