# Within group variation - low sample size

I'm looking at activity data from squirrels and want to see if my study animals follow daily rhythms. The problem is that I have a relatively low sample size:

I have data from 13 squirrels for

• time emerging from drey (hh:mm)
• time going into drey (hh:mm)
• time active (as a result of the two above; in minutes)

with 4-16 observations (= 4-16 days, the animals emerged and went into the drey only 1 time per day) per individual (mean 8.7 +- 4.4) for all 3 variables - I thought of testing all 3 variables separately.

Does anyone know about a statistical test to investigate within individual variation with this low sample size? Or is the sample size to low for statistical testing anyways? I've tried to find a solution for this but I'm struggling.

If you know about a possibility to test this in R, please let me know which package would be helpful.

• Although your stated objective focuses on "daily rhythms," your description of the data does not include any information about how many days you observed. Since a daily rhythm couldn't possibly be identified with a single day's observation, wouldn't that be important to know and account for? Assuming you do have multiple days, how do you think of and plan to characterize a "rhythm"? Were the data obtained from continuous observation or otherwise (and if so, how)?
– whuber
Jun 11, 2015 at 17:42
• sorry you are right. I just edited the post - the animals just emerged /went in 1 time per day, so this results in 4-16 days per individual (4-16 values for each of the three variables). I want to see if on the individual level, the animals emerge at approx. the same time, go in at the same time and spend similar times active per day.
– Pat
Jun 11, 2015 at 18:20