I look down a microscope at a spot on a sample and a) count the number of
features I see, N and b) measure the spot area A . From that I calculate the
feature density, D=N/A. This is repeated for a number of spots, M
I don't know anything a priori about the distribution of feature density D
across the whole sample.
My question is;
How many spots should I look at to estimate the feature density on the whole
sample and what would say a 90% confidence interval for D be?
Thanks for any pointers you can offer. Ideally I am looking for a reference
to calculate and justify the choice of the number of spots looked at.
Best Regards
Chris Wilkinson * Senior Research Scientist
CALCE Electronic Products and Systems Center
Dept. of Mech. Eng. * University of Maryland
College Park, MD 20742
Tel 301 405 4563 * Fax 253 793 7162
|