Angelo.M
Senior HTF Member
- Joined
- Aug 15, 2002
- Messages
- 4,007
It's not nearly that simple.
Sample size is selected to give a study (such as a double-blinded comparison) adequate power (which is the ability to reveal a difference between two interventions when an assumption is made that a difference exists).
The power calculation depends on the hypothesis, the variability of the outcome measure (and that's going to be very difficult to establish for, example, the performance of two power cords) and one's estimate of the relevant difference (also difficult). Very often, one needs to perform a introductory or pilot study in order to make a power determination for a larger study.
The bottom line is that a double-blind comparsion should not be performed without knowing what the power of the study will be (and, in turn, the number of subjects required to yield a statstically significant result). The only way to determine this number is to define the variables as I've laid them out.
I'm not a statistician by training, but I work with them all the time and deal with double-blinded and other similar comparative analyses every day.
