Which standard deviation to use?


I'm new here on the forum. I'm from the Netherlands and working in the field of acoustics.

I have a question, which might be quite simple. For acoustic research, we measure the reverberation time of a room by making about 10 individual measurements in a room. This is done by producing noise in a room, then turning of the speaker and then measure the time it takes for the sound to get under a certain specified level.

My question is now: to calculate the standard deviation of the reverberation time, should I use the sample SD (n-1 version) or the population SD (n-version). My idea is the sample SD, since we're only taking a few samples of the whole (infinite) population.

Last edited:
Hi, acoustics.

I agree with you that for the small-sample case in question, the biased estimator with 1/(n-1) should be used.