What are the factors that best differentiate the two groups?

#1
Hello Everyone,

I am conducting a sports study. I have these data and I would like to answer this question: What are the factors that best differentiate the two groups?
I have 2 groups of 17 professional athletes, 17 NBA and 17 MLB players with a total of 64 factors (kinetics and kinematics). They performed a lateral jump on their dominant leg then the non-dominant.
In this study, I try to find any comparison between the 2 groups. Now, I have everything I need but this. I am self-taught with statistics but I hit the wall with this one. I have tried to do a ROC but it doesn't make any sense with my values, I think..

If you could help me, that would be great.

Attached are the kinetics data only, If we can start with that.

Thank you very much for your help!
Let me know if you need more info, and sorry for my English (french aren't that good after all haha)
 

Attachments

Miner

TS Contributor
#2
It's always a good idea to graph your data to see what it is trying to tell you. Then you can decide where to take your investigation.

After looking at your data I recommend using discriminant analysis to accomplish your goal.

Sports Study.jpg
 
Last edited:
#3
It's always a good idea to graph your data to see what it is trying to tell you. Then you can decide where to take your investigation.

After looking at your data I recommend using discriminant analysis to accomplish your goal.

View attachment 2344
Thank you. Yes, I admit I have not investigated graphs so far. I've run a bunch of tests before asking this question. I first looked at what factors were significantly different between the two groups and also within each group between dominant and non-dominant lower limbs. I want to dig deeper I am sure I can still find more things I just don't know what question to ask and what stats to use. Perhaps you have an idea? Note that I only attached a couple of factors but I have many more, about 64 (kinematics and kinetics) and 34 subjects for a total of 2 176 different values that are never the same (no 1/0 or true/false I mean).

I take note about what you said, I will put my head down to understand discriminant analysis tomorrow.

Again, Much appreciated!
 

Miner

TS Contributor
#4
With that many (64) variables, you may want to consider using Factor Analysis to condense the number of variables. With that many, I would expect a lot of them to be correlated to each other.

You can run discriminant analysis on the raw data or on the Factor Scores (from Factor Analysis).
 
#5
With that many (64) variables, you may want to consider using Factor Analysis to condense the number of variables. With that many, I would expect a lot of them to be correlated to each other.

You can run discriminant analysis on the raw data or on the Factor Scores (from Factor Analysis).
Thank you for your help! Have a nice day sir.