I'm having issues with weighted averages for some data I'm working with. Here
is an example:

I have 200,000 records
Of which 8,000 have a value ranging from 3-15
The sum of those values is 40,000
The straight average is 5, based on the 8,000 records with values

Another sample
1,000,000 records
Of which 150,000 have a value ranging from 3-15
The sum of those values is 675,000
The straight average is 4.5, based on the 150,000 records with values

My question is how to make this averages weighted? The one with 200,000
records has a higher average than the 1,000,000 record sample, but I need to
reflect that based on the volume differences.

Any ideas?

Thanks.