Im doing research on the effect of salary inequality on team performance in the Major League Soccer, seasons 2007-2014. So for every season, I have data of all the salaries of all the teams, so I created some gini coefficients to measure the income inequality. For team performance/succes, the share of points of a team per season is used. But now, since I have several seasons, and the number of teams participating in the competition during every season changes, I have to take this into account. Having more teams in the competition, means that obviously, the mean share of points will be lower.

Also, for the salary inequality, I created a variable which captures the relative salary inequality: the gini coefficient of a team divided by the total of the gini coefficients of all teams, per season.

So, to see how these variables are connected, is it best to perform just a linear regression? (Using the scatter plot, I think the relationship is somewhat linear) And, for this number of teams per season problem, can this be simply accounted for by including dummies for the seasons? Because I tried it, and the coefficients of these dummies makes sense, but it seems so simple. Or should I try just a completely different modeling approach/model?

  • 1
    $\begingroup$ You could also compute a Gini coefficient of the number of points in the league. Bear in mind when interpreting that it is not possible to get the most extreme possible Gini coefficient, as each team will get some points even under "extreme" inequality. This is unlike wealth distributions, where it is theoretically possible that one persons owns everything and the rest are have-nots. $\endgroup$ – Christoph Hanck Jan 7 '16 at 17:13
  • $\begingroup$ That's a good point. I'll try that too. But what do you think about my original approach? $\endgroup$ – pk_22 Jan 7 '16 at 19:33
  • $\begingroup$ I do not see a direct issue with that, no. $\endgroup$ – Christoph Hanck Jan 7 '16 at 19:35
  • $\begingroup$ One question though, for every team I have a gini coefficient on the salary inequality. But when I compute a Gini coefficient on the number of points, I will get one value per season. This in comparison with 13 to 19 Gini coefficients of the salaries, per season. How can I model this. $\endgroup$ – pk_22 Jan 7 '16 at 19:35
  • $\begingroup$ That is true. So salary inequality refers to inequality within a team, not across teams? $\endgroup$ – Christoph Hanck Jan 7 '16 at 19:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.