Which statistics test should I run for my research project?

Question

I have a sample of 15 participants performing a "exercise screen" where 6 different exercises are scored individually.

Each participant will then get on a computerized horse that will measure pressure in reins, saddle, and leg to see if one side is weaker than the other.

Then the group will be divided in half, and half will go on to participate in a 6-week exercise program.

The entire group will meet again to retake the exercise screen of 6 different exercises and do a new ride on the simulator evaluating the three different pressure measurements.

The project is to determine whether any of the individual exercises might be able to predict weakness in any of the three areas of pressure.

There are so many variables to this project that I simply cannot narrow which statistical test would be the one to present that I will be running. Thanks for any help!

Answer 1

I realize it too late for you to do this now, but you should now for the future that the first step in ANY statistical analysis of data is to:

Step 1: Define the question, problem, etc. you are trying to answer by the analysis. Alternatively: "What the question you want to answer by doing the research, or study?

Step 2: Once you have that, you do some research of the literature to see what (if anything) has already been done to get ideas of endpoints that can be used to answer your question. You want to find an endpoint that actually measures what it's supposed to, is sensitive and reproducible.

Step 3: Only then you can be smart about the study design.

Step 4: The appropriate analysis is primarily driven by the study design.

Unfortunately, the design you describe is not simple or straight forward, and I have many questions about the dataset you're working with for your project: I need more details to provide a good recommendation.

What's you criteria for selecting 12 of the 15 people you screen?

Are there any criteria that need to be met before you screen someone?

Do you have 6 endpoints for all 3 sensors, for a total of 18 endpoints!

By the way, you might want to collect whether the subject is right or left-legged (similar to handedness, one side is likely weaker to start), so that's a variable you'd want to be able to control for.

It's unclear what the 'screening portion' is doing.

Best I can tell, you have screening data for 15 people, somehow 12 are selected & then divided into 2 groups of 6 people, 1 group does a 6 wk exercise program (details of this program?), then both groups do the same thing as they did in the 'screening' process, so you're interested in any change in muscle strength

(last (screening)measurement- 'actually' screening measurement) between the 2 groups? Or is it the sensors that you want to know about?

That's a lot of "food for thought " for you, that's all I can tell you.

Good Luck!