Time-averaged speed

27 views
Skip to first unread message

michele zanni

unread,
Feb 19, 2022, 3:52:22 AMFeb 19
to ctmm R user group
Hi Chris

I should compare the difference between two time-averaged speeds calculated with the CTMM package. Is there a function in the package that returns the t-value (or F-value) and its p-value? Or do I have to calculate it by hand using confidence interval?
In the second case I saw that there is the possibility of adding level=NULL, in function speed. This should the ensemble of mean speeds is returned. I tried but in a period of 59 days the function returned only 21 values. Why? Is this correct?

Thanks

Michele

Christen Fleming

unread,
Feb 19, 2022, 5:46:17 PMFeb 19
to ctmm R user group
Hi Michele,

In the development branch on GitHub, you can run a meta() function with variable="speed" to compare population speeds. I just checked on some fitted models and this seems to work fine for comparing individuals. You just have to package the individuals into lists, like they are samples of a population. E.g.,

meta(list(Alice=list(ALICE),Bob=list(BOB)),variable="speed")

verbose=TRUE will additionally return p-values, but I would generally recommend reporting the effect sizes foremost.

Regarding level=NULL, the ensemble here is not the set of days, but a random set of mean speeds (averaged over all days) that are consistent with the data and sufficient to meet the numerical error threshold. If the data are finely sampled, that number may be few.

Best,
Chris
Reply all
Reply to author
Forward
0 new messages