• Story Name: Great Pitchers
  • Story Topics: Sports,
  • Datafile Name: Great Pitchers
  • Methods: Regression, Polynomial Regression, Residuals, Interaction,
  • Abstract:

    The linear regression of hits, control, and clutch measures on the earned run averages (ERA) of leading baseball pitchers of the 1920-1950 era accounts for 96.7 percent of the variation on ERA's. However, a check of the residuals against the predictor variables finds a non-linearity with respect to control (CON). Intro- ducing the square of control (we suggest using standardized variables) improves the r-squared to 98.2. While this seems high, it is suggested that the student form all possible product variables (6) and all possible second degree terms (4) and run a stepwise regression of the resulting 14 variables on ERA. The result will include one product (interaction) term and one non-linear term

  • Images:
    Earned Runs

Contact Us

© 2017 Data Description, inc. All rights reserved.