• Examine the plots and tests of residuals for detection of autocorrelation and heteroscedasticity. Write more about all the heteroscedasticity tests base on the data in the Descriptive stats-1.xls. How you test it. if there is the heteroscedasticity how you fix it? Please see the example project for further help. Predicting NHL Salaries Question How can the Salaries of National Hockey League Players be predicted? Is a player’s salary dependant on the variables of years in league, games played, points, plus/minus, and position played? Dependent Variable
Y1 - 2010/2011 Salary
The 2010/2011 salary is the dependant variable that consists of the NHL player’s current salary for the 2010/2011 hockey season measured in US dollars. Independent Variables
X1 - Years in League
The years in the league measures the number of years a player has played in the NHL. This variable represents how much playing experience and seniority a player has.
X2 - Games Played
Games Played is the total number of games the player has played to date. This number can reflect a player’s value to their team based on how many games they play due to injuries or performance issues.
X3 - Position
Position is the position that a player is employed in during the game. There are a total of five skating players per team on the ice during even-strength play. These five players are divided up into two primary roles, forwards and defence. There are three forwards who provide the offence for a team and generate the majority of the points, and two defensemen who try to prevent the opposing team from scoring. This variable will be either (1) or (0). Forwards will be (1) and defence will be (0).
X4 - Points
Points are the total points that a player has scored through out their NHL career. This number is the sum of their goals and assists and is used to determine a player’s offensive capability.
X5 - Plus/minus
Plus/minus is a player’s cumulative goal differential measured through out their career. Plus/minus is either a positive number (good) or a negative number (bad). For every player that is on the ice when their team scores, they are credited with a plus one (+1), for every goal scored against their team while they are on the ice, they receive a minus 1 (-1). Plus/minus is a statistic that is used to measure a player who contributes to their team in other ways than scoring points, mainly for players who are used in defensive roles.
Expected Relationship between Variables
X1 - Years in League
A player’s experience level and seniority increases for every year they have played in the NHL. A player’s salary is expected to increase if the number of years increases. Although a player’s skill level might start to diminish after a high number of years which would result in a negative relationship to their salary.
X2 - Games Played
Hockey players are paid to play hockey. So the size of a player’s salary should be dependent on the number of games played in relation to the number of years in the league. For every game increase in games played, that player’s salary is expected to increase.
X3 - Position
A player’s position will affect the number of points and plus/minus a player has. Generally forwards have higher point totals than equally skilled defensemen and defensemen have a higher plus/minus than forwards. If a defenseman has an equal or higher point total than a forward then his salary is expected to be higher.
X4 - Points
This variable is probably the largest factor in determining a player’s salary. For every one unit increase in points, the player’s salary is expected to increase.
X5 - Plus/minus
A positive plus/minus should result in a higher salary than a negative plus/minus. For every unit increase in plus/minus, the player’s salary is expected to increase.
Data Source
NHLPA.com
The website for the National Hockey League Players Association contains the 2010/2011 salaries for all the current players in the National Hockey League. It also contains the statistics for all of the independent variables.
NHL.com
The official website for the National Hockey League also contains in-depth statistics for all of the independent variables.