1

I have done an experiment in which I have 8 repetitions per test subject (for each of several measurements). Of this I took the mean. Then I did this for 8 test subjects in total. So basically I took the mean of the repetitions per test subject and then I took the average of all test subjects. Now I'm trying to determine the statistical significance of the mean of 8 means. How would one do this?

Luc Evertzen
  • 11
  • 2
  • Your data has a hierarchical structure so hierarchical model is more appropriate in here, e.g. http://stats.stackexchange.com/questions/129783/is-this-logit-model-a-multilevel-model-and-what-is-the-correct-way-to-model-it/129839#129839 or http://stats.stackexchange.com/questions/118473/measurements-from-two-raters-should-i-use-multilevel-random-effects-model/122024#122024 – Tim Jul 07 '15 at 09:47
  • 1
    A "mean" doesn't of itself have a property of "signficance". A test statistic in respect of some specific hypothesis may do. What hypotheses are being considered? – Glen_b Jul 07 '15 at 09:59
  • @Glen_b What I have are a number of perturbation strengths applied on a test subject. I then measure their step response. I'm trying to determine whether or not the confidence intervals of the different perturbations overlap. If they do, then you can not be sure that the different perturbations actually yield different results, since their standard deviations overlap and they could theoretically give the same result. That's about how deep my understanding of statistics goes, I've never had a second of statistics in my life thanks to curriculum reform – Luc Evertzen Jul 07 '15 at 11:57

0 Answers0