The ethnicity estimate is usually the first thing people look at when their DNA results arrive. It is also, almost certainly, the result they are most likely to misunderstand. The percentage breakdown — 43% German, 22% Irish, 18% Scandinavian, 12% French — looks precise and authoritative. It is neither. Understanding what those numbers actually represent, and what they cannot tell you, is one of the most useful things a genealogical DNA tester can learn.
This is not a critique of DNA testing. The match list and segment data that come with a DNA test are powerful genealogical tools, as this series has documented across many articles. But the ethnicity estimate is a fundamentally different kind of output, and treating it with the same confidence you’d place in a match list is a mistake that leads researchers down the wrong path.
How the Percentages Are Generated
Every testing company generates ethnicity estimates using the same basic methodology: admixture analysis. Your DNA is compared against a set of reference populations — groups of living people selected because their families have lived in a particular region for multiple generations, and whose DNA therefore represents the genetic profile of that region. An algorithm examines your DNA in small windows across all 22 chromosomes, determines which reference population each window most closely resembles, and sums the results into percentage estimates.
If you have many of the same genetic markers as a Reference Population, it is assumed that you inherited these genes from that Reference Population. The company then translates the statistical output into readable categories — “German,” “Irish,” “Scandinavian” — and displays them as percentages.
The genotyping that underlies this process is highly accurate. AncestryDNA uses a microarray chip to scan over 700,000 genetic markers across your genome, with genotyping accuracy rates above 99%. That part is not in question. What is in question is the interpretation of what those markers mean in geographic terms.
The Reference Population Problem
The most fundamental limitation of ethnicity estimates is that the accuracy of your ancestry ethnicity estimate depends largely on your ancestral background and the reference populations available. A reference population represents a group of actual living individuals from a single location who’ve submitted ethnicity genetic tests to a company.
Reference populations are modern. They represent people alive today whose families have lived in a region for several generations — not the people who actually lived in that region two, five, or ten centuries ago. When an algorithm says your DNA resembles the “German” reference population, it is saying your DNA resembles people living in Germany today whose grandparents also lived there. That is a meaningful statement, but it is not the same as saying your ancestors came from Germany. The people living in Germany today are themselves the product of centuries of migration, conquest, trade, and intermarriage across central Europe.
The size and composition of a company’s reference panel varies significantly, and this matters. AncestryDNA’s reference panel skews heavily toward the Americas, Britain, and Ireland — reflecting its predominantly North American user base. MyHeritage, with a large user base in Europe, the Middle East, and Israel, has invested heavily in reference populations for Eastern Europe, the Mediterranean, the Balkans, Jewish communities, and Central Asia. For Dillman researchers with German and Alsatian roots, this difference can be meaningful: MyHeritage may distinguish between Ashkenazi Jewish ancestry and central German ancestry more precisely, and may differentiate between Rhenish and Bavarian patterns that Ancestry collapses into a single “Germanic Europe” category.
Human populations have always migrated and mixed, so drawing clean lines between “Scottish” and “Irish” or “Nigerian” and “Cameroonian” DNA is a statistical exercise, not a biological certainty. The genetic difference between a person of Scottish descent and a person of Irish descent is, from a population genetics perspective, extremely small. Both groups share most of their ancestry with the same Bronze Age farming populations that entered Britain and Ireland roughly five thousand years ago, and with the same Indo-European-speaking steppe populations that followed. The DNA does not know which side of the Irish Sea it came from.
The Time Depth Problem
Here is the central issue that most ethnicity estimate marketing quietly elides: your DNA does not reflect your family tree. It reflects your biological inheritance — which extends back not four or five generations but potentially hundreds of generations, across thousands of years of human movement across the planet.
A 2% DNA ethnicity result typically represents ancestry from approximately 6 to 8 generations back, or roughly 150–200 years ago. However, DNA inheritance is random, so the actual ancestor could be closer or more distant. But the algorithm is not actually detecting ancestry from that specific generation. It is detecting DNA patterns that happen to resemble a modern reference population — and those patterns may have been introduced into your lineage hundreds or thousands of years before your documented ancestors were born.
This has a specific implication for Dillman researchers and anyone researching central European ancestry. A result showing “52% German” does not mean that 52% of your documented ancestors came from within the borders of modern Germany. It means that 52% of your genome resembles patterns found in modern people from that region. Those patterns might have been introduced by Bronze Age migrants, by Roman-era population movements, by medieval migrations, or by the ancestors documented in your family tree. The DNA cannot distinguish between these possibilities. It sees only the patterns, not the history.
Consider what this means in practical terms: a person whose documented family tree is entirely German for ten generations back might show 40% German, 30% Scandinavian, 15% French, and 15% British on their ethnicity estimate — not because their tree is wrong, but because the ancestors of modern Germans also include the ancestors of modern Scandinavians, French, and British people. These are not separate groups at the DNA level. They are recent divergences from a shared central and northern European genetic pool, and the algorithm cannot cleanly separate them.
Why Your Results Differ Across Platforms
One of the most disorienting experiences for new DNA testers is discovering that their ethnicity percentages change significantly between platforms — or that the same test produces different results when a platform updates its algorithm.
AncestryDNA has updated their ethnicity estimates multiple times, with each revision incorporating larger and more diverse reference populations. These updates generally improve accuracy, though they can be disconcerting if you had grown attached to your previous results. Changes in your results reflect improvements in the science rather than changes in your actual DNA.
The differences between platforms are most visible in the specificity of their European categories. Ancestry uses broad labels like “Germanic Europe” and “Sweden & Denmark.” MyHeritage’s system — which it calls Genetic Groups rather than ethnicity percentages — breaks European ancestry into far more granular regional communities, sometimes identifying ancestry at the level of specific German states or Polish regions. This granularity is not necessarily more accurate; it reflects MyHeritage’s larger and more diverse European reference panel and its investment in identifying fine-grained genetic communities within Europe’s relatively homogeneous genetic landscape.
Living DNA goes further still for British Isles ancestry, distinguishing between regions within England, Scotland, and Wales with a specificity that Ancestry and MyHeritage don’t attempt. That doesn’t make Living DNA more accurate in any universal sense; it means its reference panel is better calibrated for that specific ancestry. A result that reads “43% Germanic Europe” on Ancestry might read “18% German, 12% Austrian, 8% Swiss, 5% French” on MyHeritage — not because MyHeritage knows more about your ancestors, but because it has drawn more lines within the same genetic territory.
Small Percentages Are Nearly Meaningless
Large percentages (over 20%) from major regions are typically reliable, while small trace amounts under 2% carry significant uncertainty. This is a polite way of saying that a 1% result for “Sardinian” or “Basque” or “West African” is, for most purposes, not genealogically meaningful. These small percentages are within the statistical noise of the algorithm. They may reflect genuine ancestry from that region, or they may reflect random variation in how your DNA happens to align with a reference population at a particular genomic window. Without additional evidence, there is no way to know.
Researchers who spend time trying to identify which ancestor was “the 2% Scandinavian” are not doing genealogy. They are chasing an artifact of an algorithm.
What Ethnicity Estimates Are Actually Good For
Confirming broad regional ancestry. A result showing 65% broadly European ancestry with a concentration in central European categories is consistent with documented German-heritage ancestry. It does not prove it, but it aligns with it. Large-scale patterns are meaningfully informative.
Identifying unexpected deep ancestry. A result showing significant non-European ancestry that has no place in your documented family tree is worth noting — not because the percentages are precise, but because a consistent 15–20% from a non-European category almost certainly reflects real ancestry rather than algorithmic noise.
Comparing between platforms. Running your raw DNA through multiple ethnicity systems — Ancestry, MyHeritage, Living DNA, and GEDmatch’s reference populations — and looking for consistent patterns across all of them is more informative than trusting any single platform’s output. A result that shows up consistently across multiple systems with different reference panels is more likely to reflect real ancestry than one that appears only on one platform.
Comparing between close relatives. Since siblings do not inherit exactly the same DNA, comparing ethnicity estimates between siblings can sometimes reveal patterns that neither sibling shows alone, and can help confirm or qualify specific percentages.
Providing context for match patterns. If your match list shows many people who cluster as Scandinavian or British in their ethnicity estimates, that tells you something about which ancestral line those matches likely connect through — not with certainty, but as a useful starting hypothesis.
The Right Frame of Mind
The most useful way to think about an ethnicity estimate is as a rough geographic sketch of your DNA’s population history — not a genealogical report about your documented ancestors. It answers a question about population genetics across thousands of years of human movement. It does not answer the question most people are asking, which is where their great-great-grandparents came from.
For Dillman researchers, this distinction matters particularly in interpreting central European results. The convergence of German, French, Alsatian, Swiss, Dutch, and Scandinavian categories in many Dillman family results is not a puzzle to be solved by finding hidden Swiss or Scandinavian ancestors. It is a reflection of the shared genetic history of northwestern and central Europe, in which the people who became “German” and the people who became “Scandinavian” were not genetically distinct until relatively recently — in population genetics terms, anyway.
MyHeritage’s more granular European categories may be particularly useful for Dillman researchers. Where Ancestry returns a broad “Germanic Europe” result, MyHeritage’s Genetic Groups may identify patterns more specifically associated with the Rhineland, Hesse, or Baden — the exact regions where the DFA’s Family Group Sheet database places the majority of Dillman origins. This is not a reason to trust MyHeritage’s specific numbers more than Ancestry’s, but it is a reason to upload your raw DNA to MyHeritage (free) and compare the outputs before drawing any conclusions from either platform’s results alone.
The match list is where the genealogical work lives. The ethnicity estimate is the backdrop.
Sources
ScienceInsights. “Is Ancestry DNA Accurate or Just an Estimate?” April 29, 2026. scienceinsights.org.
HCHS Online. “How to Read Your DNA Ethnicity Results.” May 6, 2026. hchsonline.org.
Genomelink. “Your Hidden Family: The Mystery in DNA Ancestry Testing.” January 29, 2025. genomelink.io.
DNA Weekly. “Best Ethnicity Estimate in 2026: Explore Your Ancestry.” March 4, 2026. dnaweekly.com.
Note: This post was researched and written with the assistance of artificial intelligence (AI). All content has been reviewed for accuracy as of August 2026. All errors are the responsibility of the author.
Discover more from Dillman Family Association
Subscribe to get the latest posts sent to your email.