A regression analysis of predictors on the productivity indices of Major League Baseball: 1985–2003
This ex post facto nonexperimental constructive replication study revisited earlier work by Grusky (1961, 1963), Gamson and Scotch (1964), and Gordon and Becker (1964) on whether Major League Baseball (MLB) managerial succession reflected scapegoating behaviors. Also there was interest in learning if selected independent variables were predictive of a team's won-loss record, termed the productivity index.
Nineteen-years of relevant data was collected from 26 of the 30 MLB teams. The four excluded were relatively new expansion teams. The dependent variable of team efficiency, a productivity index, was the won-loss records during the tenure of a specific manager with a given team. Eight independent variables we selected as predictors: on-base percentage, on-base plus slugging percentage, walks plus hits per inning pitched, stolen-base efficiency, total team salaries, length of manager tenure, average strikeouts per nine-innings, and managerial change.
The conclusions were: (1) The two most potent predictors of team efficiency were the on-base plus slugging percentage (OBS) and the walks plus hits per inning pitched (WHIP) statistics. According to the model studied, those two independent variables accounted for 26% of the variance in prediction. Using all eight independent variables resulted in a 27% variance. (2) Capable players who consistently performed up to expectations were the determining factor influencing productivity indices. (3) According to the model studied, managerial succession was not critical for improving a team's productivity index.
Recommendations on pursuing future research included: (1) Manager approaches on actions resulting in success or failure, on intangibles such as risk-taking and team culture. (2) Qualitative approaches including interviewing current and former players, managers, team management and owners on the relative role and importance of a manager. (3) Mixed method approaches including interviewing media personnel and relevant fans on the importance of a manager. (4) Repeat the study using a longer time period and use other independent variables.