The Correlations Between Salaries of NBA Players and Their On-court Performance Indicators
ZiWEI YONG, 460228934
21, FEB, 2016

Contents Abstract 2 1. Introduction and literature review 3 1. Introduction to Regression Model 3
1.1 Regression Model 3
1.2 Simple Linear Regression Model 3
1.3 Multiple Linear Regression Model 4
1.4 Multicollinearity 4
1.5 Conclusion for Multiple Linear Regression Modelling 5 2. Literature Review 5
2.1 Introduction of NBA 5
2.2 Reasons for High Average Salary in NBA 5
2.3 Salaries based on long term contract 6
2.4 Reasons for Choosing On-court Performance Data 7 3. Preparations for Running the Multiple Regression Model 8
3.1 Objectives of this paper 8
3.2 Introduction of the variables 8
3.2.1 Dependent Variable 8
3.2.2 Independent Variables 8
3.3 Data Source 9
3.4 Scattered Plots 9
3.4.1 Scattered Plots of Salary and Independent Variables 9
3.4.2 Residual Scattered Plots 10 4. Multiple linear regression modelling 11
4.1 The Adjusted R² 11
4.2 The Histogram 12
4.3 Model Generated by Analysis: 12 5. Ethical Problems 13
5.1 Sample Size 13
5.2 Data for Kobe Bryant 14
5.3 Excluded Related Independent Variables 14
5.4 Multi-collinearity 14 6. Conclusion 15

Abstract
This paper examines the correlation between NBA players’ salaries and their on-court performance indicators. Before getting into the relationship, I would introduce the essence of what is regression model and how to interpret it, then we would move on to the introduction of NBA and how this model have been wildly used in testing the relationship between NBA players salaries and indicators. In the third part of this paper, regression models tell us that: firstly, age do not affect their salaries; furthermore, as they get better score in PER (play ‘efficiently’ in game), the chance of getting longer contracts seem to be bigger; the