NHL Regression Analysis Write-Up

1. I chose the topic because I am a big hockey fan, namely the Los Angeles Kings, who I have been following for quite a while. I am trying to find out whether the amount of goals scored, the number of minutes penalized and a Canadian nationality correlates to a player's salary amount.
2. I think my model will predict estimates of the conditional expectation of my dependent variable (annual salary), given the independent variables (goals, PIM & Canadian nationality). I expect the b for x1 (goals) to be positive as the more goals one scores, it warrants higher skill, thus higher pay. I expect the b value for x2 (PIM) to be negative as an increase of penalty minutes translates into less playing time which can result in lesser opportunities to score/assist on plays. Also, players that accumulate more penalty minutes are more likely to play the role of "enforcer" than "goal scorer", thus are historically paid less than most players. I expect the b value of x3 (Canadian Nationality) to be positive as most players in the National Hockey League are of Canadian nationality.
3. The coefficient parameter for annual salary, goals, penalty minutes and nationality are as respectively follows: 39058.28, -10044.7929 and -354824.588.
B0: This intercept literally means if goals and PIM were 0 while nationality was not considered at all, that annual salary is 39058.28. That would not be sufficient. We are relatively more interested in the change of the salary rather than the value of the salary provided. Ultimately, this intercept holds no meaning
B1: For every 1 point increase in goal, there is a 39058.28*(1 point) increase in dollars of salary
B2: For every 1 point increase in PIM, there is 10044.7929*(1 point) decrease in dollars of salary
B3: If the player is of Canadian nationality, there is a 354824.588*(1 point) decrease in dollars