Ethnicity is Just an Estimate – Yes, Really!

Lots of people will have received DNA tests as gifts over the holidays. This pleases me to no end, because I know I’ll match any number of them and maybe, just maybe, those matches will help me fill in those pesky blanks in my tree or break down brick walls.

However, for the most part, those testers probably aren’t genealogists, at least not yet. They are most likely curious about “who they are” or didn’t even realize they might be curious about anything until they unwrapped that gift and discovered a DNA test inside.

Let’s hope they test with one of the major 4 companies, being Family Tree DNA, MyHeritage, Ancestry or 23andMe. (Sale prices are still in effect.) Some additional firms are certainly reputable and provide ethnicity only tests (meaning no matching), such as the Genographic Project, LivingDNA and Insitome, but then there are also a growing number of questionable pop-up DNA testing, upload sites and interpretation “services.” And yes, I’m using that word loosely. Buyer beware.

For genealogists, the gold is in the cousin matching. We already know that DNA is more than ethnicity, and ethnicity is far more than percentages.

Ethnicity, for the most part, is a shiny red bauble that the magic wand of advertising transforms from a diamond in the rough into the glittery Hope diamond with a free kilt to lederhosen conversion (or vice versa) thrown in to boot.

Ethnicity bauble

Yay – Results are Back

Everyone who received DNA test kits during the holiday season has hopefully spit or swabbed and mailed and is now waiting excitedly. Waiting is always the hardest part!

Soon, they will be discussing their ethnicity results. Reactions will vary, swinging like a pendulum – and you may well get to help interpret.

  • Some people will be thrilled because their results will confirm what they see or believe and their family stories. For example, if their family carries oral history of a Native American ancestor and their DNA ethnicity results show Native American heritage, they’ll be thrilled.
  • Some people will be pleasantly surprised with whatever information they receive – treating their ethnicity results as a nice package to unwrap, regardless of what’s inside.
  • Another large group will be confused? My mother said her grandmother was French! Why don’t I see France? Substitute <country of your choice> for French/France.
  • And then we have the truly upset. The distraught. The entirely disbelieving. “My great-grandmother was a full-blood Cherokee. Why don’t I show Native American? These tests are wrong!”
  • Some people will doubt their parentage based on ethnicity results alone. This is NOT under any circumstances appropriate. Please have them read Ethnicity and Physical Features are NOT Accurate Predictors of Parentage or Heritage.


To help people understand, you may need to explain about how Native Americans, especially east of the Mississippi were admixed very early in our national history, so their “fully Native” ancestor probably wasn’t.

You can explain about how autosomal DNA is diluted in each generation since their Native (or French, or Italian, etc.) ancestor lived – to the point that the Native DNA might not show today.

You can talk about reference populations, or the lack thereof, and that people in France and Israel can’t legally take DNA tests for recreational purposes.

You can educate people about how we all need to research our genealogy, and how, as Blaine Bettinger writes in this classic article, we have both a genetic and genealogical tree. The ancestors are always there in our tree, but we may not have inherited measurable DNA from a particular individual if they are several generations back in time.

If that coveted Native ancestor doesn’t appear in their DNA, then they need to look in their family tree. She or he might be waiting there, AND, they may still be able to prove their Native heritage using either Y or mitochondrial DNA testing at Family Tree DNA.

There’s more than one kind of DNA and more than one way to prove Native heritage.

The Underlying Truth

But the truth of the matter is, while each and every one of those statements above is entirely valid, the fundamental truth about ethnicity testing is that…

Ethnicity percentages.png

Yes, really.

Let’s take a look at some of the reasons why.

Size Matters

Everyone in the Americas (except for Native American, First Nations or aboriginal peoples) wants to know where their ancestors “came from.” As genealogists, we deal with no records, damaged records, misplaced records, burned records, rapid westward migration with no links “back home” and at least three wars on our soil. It’s no wonder that we often can’t track those ancestors back across the pond or even to the shore.

Therefore, we hope that DNA testing can help us bridge that gap. And indeed, both Y and mitochondrial DNA testing is wonderful for doing just that for matrilineal and patrilineal lines.

But ethnicity results, in most cases, are really only useful for making continental-level discoveries. What we really want, refinement and granularity to the country level within Europe, for example, isn’t really feasible.

Size is part of the reason why. Look at the size of the contiguous 48 US states as compared to Europe, courtesy

Ethnicity US over Europe.png

Would you expect to be able to tell the genetic difference between people that live in Washington State from people that live in Idaho? That’s roughly the same distance as from the UK to Germany. France is located down in California and Nevada.

Can you tell the difference genetically between people who live in Washington State from California or Nevada? That idea sounds rather preposterous when you look at it that way. Now, is it any wonder that your ancestor’s “French” doesn’t show up, but German does?

Ethnicity Texas over Europe

Here’s Texas compared to Europe. Can you tell the people in Dallas from the people who live in San Antonio from the people who live in Houston, genetically? That’s the same difference as Germany, Italy and Austria. The Czech Republic is over near Shreveport. You get the drift.

Western European Countries are the Size of US States

Western European countries are even more difficult.

Ethnicity states over Europe

How about discerning the difference between Indiana and Illinois residents, or Illinois and Missouri? European countries are the size of medium sized US states. Larger states, like Texas cover most of the Iberian Peninsula including Spain and Portugal and reach over into Morocco.

To make this relatively small region even more complex, people have moved freely across these areas for thousands of years. The people from the Russian Steppes moved into Eastern Europe displacing and assimilating with the hunter-gatherer population that had resided there for millennia.

The Germanic tribes moved towards the coast and into the British Isles. The people from “Indiana and Ohio” moved into “Illinois” and then that entire group populated parts of Scandinavia. According to a recent genetic paper, some of those “New Yorkers” and on east moved into Scandinavia too.

Oh, and the Sephardic Jewish people moved from the Middle East into “Texas” aka Spain and then on up to “Indiana, Ohio and Pennsylvania” some 500 years ago to join their Ashkenazi brethren. Fortunately, Jewish people generally stayed together and didn’t intermarry or assimilate much into the local population, so we can still identify them genetically.

Europe is indeed a great melting pot.

Ethnicity Alaska and states over Europe and Asia

Adding the largest US state, Alaska onto the map makes the rest of the states and their corresponding European countries look really tiny.

Ethnicity is Really Only Reliable at a Continental Level

Ethnicity really is only reliable at a continental level, plus Jewish and in particular, Ashkenazi. Very small or trace percentages may not be reliable at all. We’ll discuss ways to prove or disprove minority admixture in my next article, Minority Ethnicity Percentages – True or False?.

This continental-level-only phenomenon is more understandable if you look at a world map.

Ethnicity continents

It’s extremely difficult to discern any reliable level of granularity between regions as tiny as US states in Europe, no matter how badly testers want to know. Of course, that doesn’t keep the testing companies from trying, and kudos to them. As they make improvements, your intra-continental estimates will change over time – so don’t fall in love with them. And don’t trade that lederhosen for a kilt or vice versa – or get that Viking tattoo just yet.

It’s much more reasonable to rely on ethnicity estimates based on much larger regions, where people after migration have been separated from people in the other regions for a much longer period of time, allowing time for unique mutations to develop.

Less admixture happens with greater geographic distance. People who aren’t neighborly don’t produce offspring because begetting requires proximity. Mutations that occurred after the populations split into different regions are found only in the new or the old populations, but not both – at least not in high frequencies. Of course, population boundaries are fluid and people (continue to) move from place to place, back and forth.

What You Can Do!

When your family and friends begin to discuss their confusion or disappointment with their ethnicity results, you’ll have this article to explain the situation visually. Please feel free to share and encourage them to learn more.

Sometimes it’s difficult to be the cold voice of reason in a positive way, but there is so much more to learn. I always hope to spark curiosity about why, and then provide ways that the person can fall in love with discovering their ancestors and ancestry.

Another good resource is the article, Ethnicity Testing – A Conundrum which explains how DNA ethnicity testing actually works – in terms that everyone can understand.

If your family is wondering what happened to their Native American DNA, you’re not alone. I’ve put together a page of Native American Resources to help everyone!

Have fun, enjoy and let’s hope that newly baptized ethnicity testers will like the water enough to engage in a bit of genealogy. You can encourage them by helping construct their first tree by recording what they know about their parents and grandparents. Maybe give them a taste of success by helping them find a record or two. Give them a taste of genealogy crack.

You never know, it just might be habit forming!



I receive a small contribution when you click on some (but not all) of the links to vendors in my articles. This does NOT increase the price you pay but helps me to keep the lights on and this informational blog free for everyone. Please click on the links in the articles or to the vendors below if you are purchasing products or DNA testing.

Thank you so much.

DNA Purchases and Free Transfers

Genealogy Services

Genealogy Research

46 thoughts on “Ethnicity is Just an Estimate – Yes, Really!

  1. You are one of few who tell those who purchase DNA tests what to expect from their results. Certainly none of the four major companies make this clear. Since testing, I’ve transferred my results to the major four companies. FTDNA does the worst job being user friendly for novices such as myself. Ancestry and My Heritage do the best job in helping new and older customers to be encouraged through sending emails announcing new matches, particularly finds with which we share the most dna. There may be small print somewhere that all of the companies explain what to expect, but in their advertising it is not noticeable. It’s time consuming and most people have full time jobs and must be enthused enough to understand the smallest of information about genealogy. So it’s unlikely that consumers will get the important news that they are not going to get the ethnicity they were thinking. There are many experts in this field but many, many more who are only hearing and seeing the ads and lowered prices promising to tell us our roots and ethnicity. More and more disappointed consumers are sharing the truth with those who ask them what they think. I think those who “sell” their tests owe more to the public than results they can’t deliver and reduced prices.

  2. I am beyond weary of hearing whining and complaining about ethnicity estimates from those people who expected a “kilt to lederhosen conversion”. They have no idea how to read the results and worse yet, don’t want to. They just want their pedigree handed to them. They don’t want to believe they are working at ethnicity backwards and if they really wanted to know, start researching. Ancestry has been the biggest problem for several years. Their add campaigns are working for them. Not so much for the rest of us who are researching our family.

  3. I have my DNA with several labs, each one is ‘somewhat’ different. However, at, they seem to have really watered it down. I went from the several ethnicity results they have had over the years, to a common Ashkenazim and General European. Completely removing my Irish, Scottish, Welch, to “replace it with “British”, which they never had me with any ‘English’ since 2011. I am corresponding with known Turner’s from my tree, to find out what their results are to help me see the bigger picture. Seems like a fun challenge! :-).

  4. I enjoyed reading your article comparing the size of European countries with the USA states. I have a cousin in London who rode his motorcycle all the way to Rome and back home. it also is easy to get around Europe by train. Big question – I wonder how people traveled long distances ages ago without cars and planes.

  5. Love, love, love this article, Roberta!! Europe was so tiny! (you’ve shown us now!!) Even though only the affluent traveled to other countries back in the day, many of us who are several generations’ worth of products of marriages of young women and men who apparently didn’t listen to their parents about who they should marry and who they shouldn’t, should realize even if we’re from colonial lines in the USA–wow–there have been thousands of years of folks traveling around and having babies, no matter how many generations back: and a lot of times no one wrote anything down!! I have been to at least one holiday gathering where I’ve been tempted to throw up my hands about trying to explain all these promises of finding our “ethnicity.” For heaven’s sake! It’s an estimate! “Estimate!” 🙂

  6. The map overlays really aren’t entirely fair.

    The United States was colonized by Europeans in the last 300 years and much of it even more recently and has a continuous history of immigration from diverse places. The entire continental USA has only a bit more population structure than the U.K.

    In contrast, with some important exceptions, European subregional gene pools were more or less in place in the Bronze Age (by ca. 3500 years ago) and you can count the number of significant mass migrations in that entire time period on your fingers. Sardinia’s gene pool was mostly fixed by about 7000 years ago (the closest match to Europe’s first farmers).

    There has also been more admixture between originally distinct populations in the U.S. in the last century than there has been in the last thousand years in Europe. You can genetically distinguish Tuscany from Sicily which have lots of population structure in a way that you cannot distinguish roughly comparable Northern v. Southern California whose residents mostly have recent common origins and are admixed between subpopulations at similar rates.

    River basins and plains tend to be fairly homogeneous over fairly large regional areas (which often don’t correspond to modern sovereign state boundaries), but in highlands of Europe and on islands there is a great deal of very old population structure. Finland has more internal population structure than Germany and France combined.

  7. Roberta, I love the way you explained ethnicity estimates. For those who don’t know much about the details of European history (as well as other parts of the world) don’t realize how impossible it is to link an ethnicity to a particular country. I have a related question, though, about the control populations that each company uses to determine results. My maternal line is British Isles with one Scandinavian line and one Dutch line. My paternal line is Eastern Europe back to my 3X great grandparents. Why do the ethnicity results vary so wildly between companies – one says 4% British and 96% East European. The second says 44% British and 23% East European. The third says 9% British and 47% East European. These are way more general categories than simply English, German, Slovak, etc. In my case, company 2 at 44% and 23% is probably most accurately tied to my paper trail. Are control populations so small at each company that results would vary as widely as mine? Thanks.

    • How each company determines ethnicity is proprietary information. All companies use publicly available reference populations, such as the 1000 Genomes project. Additionally, I know that some mine their own databases. 23andMe has a page that discusses how many samples they have from various locations. I would presume the other companies do the same thing. Then, they determine the estimate using their own algorithm that they have developed. Why do they vary? I don’t know and neither does anyone else.

      • I knew the info was proprietary. If everyone knew how many people were in ethnic samples – possibly fairly small numbers – they would probably better understand how the numbers were determined. I figure the company that said I was 96% Eastern European must have had a skewed number of other Eastern Europeans in their control group. 🙂 Thank you for the clarification.

  8. Thanks Roberta. My ethnicity result picked up ‘15% Iberian’, which was unexpected as my last Iberian born ancestor was 7 generations back (actually Castres in France but closely linked to Barcelona). My wife’s result came back 100% South Asian which was also unexpected. She is Indian but from Goa and her great grandfather was a Portuguese doctor. He married a Goan but his daughter looked distinctly Portuguese. who was It will make us look more closely at the gt grandfather – I’m now wondering if he was Indian but educated in Portugal.

  9. I got a kick out of your comparison. I grew up in Washington and live in Idaho now. Culturally, they’re a lot of differences between the states. As far as DNA… not so much. Looking forward to the day when they can unlock even more information from our DNA to give us clues about our ancestors.

  10. A fine article, one again, thanks Roberta! One curiosity with ethnicity / admixture estimates from the Finnish point of view is that quite many Finns or people with Finnish ancestry get hints of Native American and / or Oceanian ethnicity. Obviously that is due to our Asian deep ancestry, not Native American or Oceanic ancestors.

    Appr. 60 % of Finnish men have Y-DNA N haplogroup and the migration route of that haplogroup is really long, via East Asia etc. hat fact – together with our population peculiar founding effects and bottle necks – might have brought certain amounts of autosomal etc DNA from those ancient times and Asian regions that is still visible in our “DNA pool”. That means Native American and Oceanic reference clusters contain rare mutations etc that come from our common ancestors who lived in Asia. Gen. gen. companies won’t take those parts away from the Native American or Oceanic just because hints can be found in Finnish cluster as well.

    This means that e.g. many Finns have common DNA with the ancient “Montana kid” in GedMatch. And some of us have dry earwax, a recessive mutation that is typical for East Asians and Native Americans. I know all of my grandparents have had that mutation as also my parents have got dry earwax.

    So with ethnicity / admixture estimates even continents are not always chrystal clear. Some hints of Native American and / or Oceanian deep ancestry (esp. for Finns) might actually be hints of East Asian deep ancestry.


    Worth of reading: Lazaridis et al: Ancient human genomes suggest three ancestral populations for present-day Europeans, Nature 2014.

    In the article: “most present Europeans derive from at least three highly differentiated populations: West European Hunter-Gatherers (WHG), who contributed ancestry to all Europeans but not to Near Easterners; Ancient North Eurasians (ANE) related to Upper Paleolithic Siberians, who contributed to both Europeans and Near Easterners; and Early European Farmers (EEF).”

    In the annex: “The set R includes all populations identified in both SI14 and SI17 as compatible with being derived from the same 3 ancestral populations, and excludes Sicilians, Maltese, Ashkenazi Jews, Finnish, Russians and Mordovians which have evidence of additional complex history.”

  11. The overlay of maps is suggestive to a point, but it homogenizes things which often still are separate. I mention only two things which keep people more confined and less sprawling: language and religion. Sure, people can still move if the incentive is great enough, but it’s not as simple as moving from Illinois to Missouri. Geography plus language plus religion when overlapping go a long way to adding up to an ethnic group.

    • The Mennonites moved from Switzerland to Germany to Holland to Russia and to the US. And that’s just a group we know about. During the 30 years was entire deaths of Europe were depopulated and later repopulated by people from other locations. People consistently move.

    • One major barrier was the Danube and Rhine rivers. We took a cruise a few years ago. On the western side, the Romans ruled while in the eastern side, Germanic tribes and barbarians ruled. The Romans couldn’t beat the barbarians on the hillier and more densely wooded eastern side.

  12. More sense than 5 years’ worth on the subject from Ancestry.
    Although they did show that 1)ethnicity varies a lot 2)how little they think the test is worth – with their ad a couple of years ago with the three sisters who test just to discuss how Irish they are scored as over their next coffee together.

  13. How would you explain that one dna company shows Jewish and matches me to a Jewish population in their database, which is also confirmed by taking a mtdna test that shows Sephardic Jew and Palestinian, but two other dna companies do not show any Jewish percentages, yet one of them gives me matches to Jewish cousins with J and K haplogroups?

    Is it because of each companies different methods of testing and if so, why are they not the same?

      • I can’t answer the technical questions about the different interpretations in the DNA tests, but I can offer some insight into the Sephardic Jews. During the Spanish Inquisition, the Jews were ordered to convert, flee, or be killed. Many chose to convert and intermarried with non-Jews. Others married Ashkenazi Jews. My mtDNA is Sephardic Jewish, yet my family history is Ashkenazi Jewish. Interestingly, many of my mtDNA matches aren’t Jewish.

        • Sometimes people misinterpret mito results, especially at the HVR1/HVR2 levels. Some people who aren’t Jewish will have some Jewish matches, especially at low levels, simply do to the common haplogroup.

        • Are you so sure they aren’t Jewish though? My family doesn’t immediately register in every test or calculator as Jewish, but we still are – just mixed with a lot of other things.

          I’ve had a lot of bad luck with Ashkenazi cousins on Sephardi segments – you most likely did have a Sephardi ancestor who just never spoke about being such; there’s a lot of reasons why they wouldn’t have. It’s more common than you might think.

          • EDIT: I mis-read, my apologies, I was thinking you meant autosomal matches and completely skipped the fact you mentioned mitochondrial. Oops!

  14. Roberta,

    While you were in Norway, did you happen to hear of any possible reference panel additions to the MyHeritage ethnicity calculator?

  15. Thanks.

    I could have sworn that I read something from MyHeritage about adding new ethnicity group samples to their reference panel later this year (2018). This would have been around January-February of 2018. Perhaps they will do so in the near future.

    • They may be. I don’t recall, and I don’t recall specifically about Norway. There was a lot to absorb and I was incredibly jet-lagged too.

      • Well, I can only imagine how busy you must have been! However, I have a feeling that you caught the announcement about MyHeritage possibly offering an option to use DNA from stamps and envelopes…how exciting!

  16. I notice that a lot of people use their surnames interchangeably with ethnicity. That’s where the trouble often starts. They ignore 7 to 8 generations of say English maternal ancestry all for the sake of the paternal line surnames. Had to explain this to my wife whose surname is Altenhoff.

  17. Very cool article. What surprised me most about my results was that I didn’t get any German at all. My grandmother was from the Conrad family. So I thought I would at least get some German from her family, From what I was told Conrad is a very old German family.

  18. Loved the article 🙂

    Three comments: as someone who’s of mixed Jewish descent, both Sephardi and Ashkenazi, and is also part of a DNA project for this heritage I have to say that the estimates for people of mixed gentile and Jewish heritage is dubious at best. Ancestry pegged us but 23&me was abysmal, both for us and for a lot of our relatives who’re directly of MENASA descent instead of admixed. Interfaith marriages in the 1800s isn’t terribly uncommon either, at least not in my family’s case, as we had Jews marrying Tatars marrying Armenian-Kurds marrying Poles. Another thing to keep in mind is that there are many groups of Jews who did not survive past the Shoah to be able to reproduce and have their descendants become part of a sample population, meaning there’s a chance there are people who should be getting pegged as Jewish that aren’t, because they can’t recognize the specific SNPs. 23&me tried to label our Tatar and Jewish DNA as Romanian, for example, and they were right in that our line goes into Romania at some point, but we are most definitely Polish Ashkenazim descended, not Romanian.

    Another thing is that I have to often explain to Sephardi relatives that they will not show up as Jewish on a DNA test – they’re often very upset and concerned and I have to go through the steps to explain to them how our ancestors are virtually indistinct from our non-Jewish neighbors, and how often we would intermarry. In my family’s case, our line came to America from then Syria now Lebanon, directly from a Maronite town, despite us being ethnically Jewish, and we have many, many Palestinian, Jordanian, and Lebanese Christian cousins because of this.

    Thirdly, it is really funny how different tests will label different things. My grandmother’s mother had indigenous Siberian ancestry, most likely Koryak or Chuckchi, and 23&me tried to claim it was Japanese (which – they were half right, the segment it pegged as Japanese we DO share with a Japanese relative, but it’s because of intermarriage between native Ainu and Koryak peoples due to geographic proximity, not that we’re Japanese). Ancestry labeled it Northwest Russia / Finland. Due to the intermarriage of Arctic people we have a few cousins in Nunavut, even. I wish people talked more about the relationships between indigenous peoples prior to European contact, genetic flow / intermarriage, and how we traveled.

  19. Pingback: DNA Results – First Glances at Ethnicity and Matching! | DNAeXplained – Genetic Genealogy

  20. Pingback: 4 Kinds of DNA for Genetic Genealogy | DNAeXplained – Genetic Genealogy

  21. Pingback: First Steps When Your DNA Results are Ready – Sticking Your Toe in the Genealogy Water | DNAeXplained – Genetic Genealogy

  22. Pingback: Native American & Minority Ancestors Identified Using DNAPainter Plus Ethnicity Segments | DNAeXplained – Genetic Genealogy

  23. German was more of a Language than an ‘ethnicity’ until the 19th century when Von Bismarck started uniting the german-speaking peoples into an Empire – for instance, German was spoken to the borders of Holland and Denmark, across Prussia and thru Poland all the way to Transylvania and back thru Switzerland to the Alsace! Intermarriage is a fact and people move for various reasons. As wars broke out migrations occurred on a grander scale, and all over Europe (not to mention the americas) – so that’s why my Hanoverian ancestors from1850 show up as western slavic, or maybe even ‘anglo’ saxon rather than middle-saxon. You’re right, it’s sort of a guess.

  24. This article was sent to me as a link from Family Tree DNA in response to a question I submitted. My daughter and I both did the autosomal Family Finder test and our ethnicities came back almost completely different, not the percentages but the categories themselves. Her mother is of a different race and that was correct per the test. The test also correctly identified us as parent/child. Both of our lists of relatives are the same, and yet again, comparing her ethnicities to those on her list of relatives, they are completely different whereas mine are basically the same as theirs. The only overlap is her 14% of Western Europe to my 94% with a further breakdown of her 4% of England, Wales and Scotland to my 49%. Everything else is different. Is this possible between only one generation?

    • If you’re talking about portions of Europe, it’s extremely difficult to tell the difference. Think of it like trying to tell the difference between Indiana and Illinois.

Leave a Reply