Which way to calculate mean is correct?

its late and my head hurts, and im sure ill kick myself in the morning for asking this. Say i have the following data:

counts  area   density
10      5      2
12      3      4

where 'counts' and 'area' are measured and are the two rows are valid replicates. density=counts/area

is the best single measure of density the mean of 2 and 4 , or does it make sense to calculate a total area 8 and a total counts 22 for a density of 22/8???

Answer 1

[Walter Roberson]

Or even sum(counts .* area) ./ sum(counts)

It depends on what is being represented. If it is a histogram type count, 10 occurrences of area 5, then you need to do the calculation the way I show above. But for some other things, sum(counts) / sum(area) would be appropriate. The mean of the densities is not likely to be correct.