Hi everyone, this is the first thread I'm opening and for sure I'm not an eagle in statistics. I'm analyzing data coming from a clinical trial I'm about to publish, but I'm facing some doubts.
In this particular situation, I'm analyzing data coming from diabetic patients in which I'm comparing the standard deviation of blood glucose curves at different time points.
Blood glucose curves are composed of glucose concentration values that are measured consecutively every 2 hours during the day. To evaluate the variation of those values we use standard deviation (SD). I'm evaluating SD in 8 blood glucose curves performed weekly (in 8 weeks) in 9 different patients.
Here is the question:
If I want to evaluate if the standard deviation changes significantly during those 8 weeks which test would be better to use if my data are not normally distributed? Friedman test would be good?
And if I'm missing some of the data, the software doesn't allow me to use Friedman test? Which other tests could I use?
In this particular situation, I'm analyzing data coming from diabetic patients in which I'm comparing the standard deviation of blood glucose curves at different time points.
Blood glucose curves are composed of glucose concentration values that are measured consecutively every 2 hours during the day. To evaluate the variation of those values we use standard deviation (SD). I'm evaluating SD in 8 blood glucose curves performed weekly (in 8 weeks) in 9 different patients.
Here is the question:
If I want to evaluate if the standard deviation changes significantly during those 8 weeks which test would be better to use if my data are not normally distributed? Friedman test would be good?
And if I'm missing some of the data, the software doesn't allow me to use Friedman test? Which other tests could I use?
