JCMathsCorrelation and regression10–15 min
Best-Fit Sniper
Line up the regression line of y on x by shrinking the sum of squared residuals.
The experiment opens when it comes into view.
Theory and model assumptions
Least-squares line of y on x: b = Sxy/Sxx, a = ȳ − b x̄, minimising Σ(y − a − bx)². r = Sxy/√(Sxx Syy) measures linear correlation.
Teacher guidance
Expected time: 10–15 minutes.