Ethnicity becomes less accurate the broader the gene pool

As a follow-up to my last entry on the improved ethnicity algorithms of Ancestry.com, I also looked at the two generations below my siblings and first cousins, all of whom are about 50 percent Italian, verified by extensive genealogical research, and now also by DNA testing. I’ve noticed, though, that in earlier updates, the amount of DNA identified of Italian origins dropped off sharply for further generations. In the October 2025 update, I see some improvement in the generation just below mine.

Three of my first cousins once removed have tested. By traditional informal methods, they would be considered 25 percent Italian, or maybe since all of us who consider ourselves half Italian tested from 40-50 percent Italian, the next generation down should come in from 20-25 percent Italian. Indeed, SL and JS show 22 percent, a definite improvement over previous results. However, FB shows only 9 percent, far below what might be expected.

Yes, I understand that genetics is complicated, and the distribution of DNA isn’t even from one’s parents and grandparents. One helpful illustration I’ve heard is that we all may come from the same batter made for chocolate chip cookies, but the number and placement of the chips will vary in each cookie. That could be a partial explanation of the discrepancy between the three cousins.

However, in the next generation down, the formula Ancestry uses is definitely problematic. I have at least three first cousins twice removed who have tested. Following the paper trail, they could be considered from 10-12.5 percent Italian. However, SL and DC came in at 3 percent, and ML at zero. It seems pretty unlikely that if you broke one of the parent cookies in half, the resulting half cookie would have no chocolate chips at all. Obviously, it gets very difficult to make ethnicity predictions when one’s gene pool becomes an admixture of eight different great grandparents, each with different ethnicities. I’m not really criticizing Ancestry for this issue, because I know it must be incredibly complicated to accurately figure out the ethnicity of people of such diverse ancestry. Instead, I just want to warn people that their results may not be as accurate as they are led to believe. I know some people who have been disappointed when their DNA results don’t match what they believe to be their ethnicity. In such cases, it is better to trust your traditional research, tracing each ancestor back as far as possible. One should still consider the DNA results, but it’s advisable to treat them with a healthy dose of skepticism, research and reality.