I currently have an exercise protocol which has high inter and intra-individual variability.
I am comparing a treatment to a placebo. Due to the high variance is it possible for me to control this when doing my analysis to see if there are any significant differences between treatments?
I don’t think that I will see any significant differences if I don’t control for this variance.
Would this be done as an ANCOVA?
Are there any other statistical tests which I could run?
I general select as the primary efficacy parametr variable with high inter and intra-individual variability is not recommended.
First step is to communicate with clinical expect for the diagnosis of this trial if really does not exist better measure of efficacy.
If not than the important part of the protocol is the sample size estimation. If effect size is fixed than the sample size is a function of the square of standard deviation. So I expect that sample size of such trial will be pretty high.
Using of ANCOVA well solve issue of high inter-individual variability ( = variability among patients), but it is not big help in case of high intra-individual variability ( = variability of repeated patients measurements ... if baseline value will be taken next day or week/month - it could be very different)
Could I ask that diagnosis of this trial?
BR Lada
ANCOVA will only control for variance if that variance is systematically predicted by a covariate.
If you have a small effect and a lot of variance, then to detect that effect you to use a large sample and/or you need to reduce the variance. There's not really any way around that.
The variance was going to be samller if Random assignments for the treatment and placebo groups.
Next step, yes you can use the statistical control such as ANCOVA, assuming that you have measured the covariate
Hi,
Thanks for replies.
I am trying to measure fat oxidation rates during exercise. Thus I wanted to compare multiple trials for fat oxidation rates both within the same person and also between individuals.
The trials are a week apart.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Statistical Software
- SPSS