I am an amateur researcher. I would like to know for the following data set, which statistical test should I use?

one compression garment and a control garment was tested among ten participants for its potential benefit towards recovery. Recovery was measured using Heart rate (HR) and Blood lactate (BLa) before and after a maximal fatigue event (exercise to failure) was completed by all participants.

Results were obtained for both variables

  1. Before test
  2. Immediately after test
  3. 20 minutes after test
  4. 40 minutes after test
  5. 60 minutes after test
  6. 12 hours after test

What is the ideal statistical model that should be applied to investigate if garment has better recovery potential than control.

Added info as requested

1.Data - Heart rate (continuous dependent variable) - measured using wearable - Blood lactate (continuous dependent variable) - analyzed using blood sample and portable analyzer

  1. Each participant was measured multiple times during the following scenarios

    • Immediately Before exercise
    • Immediately after exercise
    • 20 minutes after exercise
    • 40 minutes after exercise
    • 60 minutes after exercise
    • 12 hours after exercise
  2. Research question

    • Does garment A (experimental) produce better recovery in terms of heart rate and blood lactic acid levels than control (garment B)

Thank you

  • $\begingroup$ You should edit your question and include more information 1. Describe your data, what variables do you have and what type they are. 2. Did you measure each participant multiple times? 3. What is your research question? $\endgroup$ Commented Apr 1, 2020 at 7:04
  • $\begingroup$ What specifically is your hypothesis? As you intend to write it in your research? $\endgroup$
    – ajax2112
    Commented Apr 1, 2020 at 8:48
  • $\begingroup$ For the research question as stated, it should probably be more specific so you can relate it to your measurements. What is 'better recovery'? A lower heart rate? Lower 'blood lactate'? After 20 minutes, or after 12 hours? Or if you want to combine all these measurements, do you expect a specific pattern over time; a logistic return to normal or something? $\endgroup$
    – Gijs
    Commented Apr 1, 2020 at 9:02
  • $\begingroup$ expected better recovery is a heart rate and blood lactate value for garment A less than garment B at 12 hours $\endgroup$
    – SAT
    Commented Apr 1, 2020 at 9:13

1 Answer 1


As an amateur researcher you've chosen a pretty hard to analyze research design.

Firstly, I advise you to not do statistical testing with only 10 subjects please. Make it a case study for 10 people because 10 just isn't enough for this kind of research.

However, let's say you had more subject. What test should you use?

Because you're measuring subjects multiple times you need to correct for the correlation within one subject in that regard but because you're also using the same subjects multiple times you need to correct for correlations there as well.

The test that is appropriate for this is multilevel analyses. In multilevel analyses you define your levels to account for correlations like in your research.

To build up your understanding let's say you have a experimental design with 10 subjects in condition A and 10 in condition B and you measure them over time. Here you'll have 1 level because you only account for correlations within subjects.

                   Condition A:                             Condition B:                     

         subject 1             subject2                      subject 8      ......

T1       T2      T3       T1       T2      T3            T1       T2      T3

Now your design includes the subjects in both conditions. Here you will have on the lowest level condition and above that the subjects subjects:

             subject 1                             subject 2                                              

   A                          B             A                    B          

T1, T2, T3                T1, T2, T3    T1, T2, T3          T1, T2, T3 

I know multilevel analyses could be hard to understand at first but look at it as a glorified repeated measures anova.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.