Standard Deviation in One Direction

andrewcheong

I need to calculate the average time of an event, and I'd like to calculate standard deviation as well. The problem is - there is no such thing as negative time - zero is a boundary - so how does it make sense if the average is 4000us (microseconds) and the stdev is 5000us?

Is there a different measure that I should be using when there is a boundary on one-side, like some sort of a one-directional standard deviation? "Semi-infinite interval" comes to mind, and I Wikipedia'ed such probability distributions for more information, but I'm not sure what to make of it.

Thanks in advance!

LCKurtz

Do you have a distribution in mind that actually does that? Sometimes if the distribution is badly skewed the interquartile range gives a better description than the standard deviation.