Significance above background.

Ish

New Member
#1
Just a very brief question about determining significance from a very limited 'sample'.

A single measurement is taken at each of 35 sites. Many measurements fall within a lower limit with 2 or 3 moderately high or very high readings. How can I determine whether those readings are significantly higher than the rest of the readings or not?

I feel kind of embarrassed asking, but I don't see how traditional significance tests apply to this situation.
 

JohnM

TS Contributor
#2
A box-and-whisker plot should help in identifying which points are potential outliers. You would then need to do further investigation and/or repeated studies (and hypothesis tests) to see if those readings were bona-fide and truly much larger than expected.