# Question about regression approach for this data.

#### econkid

##### New Member
I'm working on a research project for my sports economics class, and am having some difficulties with my data. I'm a little weak with statistics and am unsure how to approach the regression. As in what should be my independent variable/ dependent variables, which to isolate in excel. Essentially what i'm trying to see is if player performance is effected by contract negotiations so I am analyzing the performance immediately before and after contract negotiations and regressing that performance (I suppose this would be my "Y"). The contract year in my data is highlighted, as in the highlighted year is the performance after the off season contract negotiation and the cell before the highlighted one is the year immediately before the new contract. Any guidance would be greatly appreciated. Let me know if anything is unclear and I will follow up. Data is attached.

#### Attachments

• 527.3 KB Views: 7

#### j58

##### Active Member
So, your hypothesis is that performance immediately after contract negotiation will systematically differ from performance immediately before? Then I don't see why you would formulate this as a regression problem. Why not just, for each player, subtract the pre-negotiation measure from the post-negotiation measure, average these across players, and test whether this average differs from 0?

#### econkid

##### New Member
Ok, so you're saying I should do a one/two tailed test to check if there is a significant statistical difference. But I also want to see how strong that difference is, wouldn't I need to have a regression for that? I am trying to use a similar analysis to David Berri and Anthony Krautmann when they did a similar analysis for the NBA. A screenshot of their model equation and summary statistics table is attached. It looks like they did an analysis that you recommended as well as a some type of regression. Perhaps my model is too simple? Should I account for other variables like Berri, to account for omitted variable bias?

#### Attachments

• 241.5 KB Views: 5
Last edited:

#### econkid

##### New Member
Disregard that previous post, actually I think I'm trying to use this regression model that was created for the a similar analysis for the NBA for my data that I've created on the NFL. Attached is the regression model. Is it possible to use my data for this model?

#### Attachments

• 147.9 KB Views: 3