As I understand it .....
The population test samples are relatively small in number. These samples have been chosen as each person in that group has been able to trace their family tree using the traditional paper trail to a specific area.
Their results are then used as a marker against which our results are compared. Each company uses different samples, hence the differing results.
I am guessing that might also explain why some of us don’t have the percentages we might expect from particular areas of the world - simply, not enough comparative and proven family trees?
Something I wonder about is, surely it would only take a couple of those within a group (particularly noticeable in the sample groups with smaller numbers) to have an NPE or two, or three, to skew the results?
No matter how thoroughly researched your family tree is, this is always a possibility ..... another reason not to put too much emphasis on the ethnicity percentages.
I hope that interpretation is correct, but I am open to being corrected.
