The first thing most people open after a DNA test is the ethnicity estimate, and the first reaction to an Austrian family’s result is often surprise. Where is the Germanic share? Why is there a Balkan percentage, or a Scandinavian one? This post explains what ethnicity estimates are, why they behave this way and how to read them without drawing conclusions the science does not support. It is not a statement about any person’s heritage.
What an estimate is
An ethnicity estimate is a statistical comparison between your DNA and a set of reference samples chosen by the company. The company decides how many regions to define, which reference people represent each, and how to draw the boundaries. Different companies make different choices, so the same person gets different results at different companies, and the same company can change a person’s results when it updates its method. The result is an estimate of resemblance, not a certificate of origin.
The genealogist Judy Russell documented a vivid case in 2020. Five siblings who tested at one company saw a large Scandinavian share appear after an update, ranging among them from about a quarter to about two thirds, while other companies gave much smaller Scandinavian figures for the same person. Her write-up, Eye of newt, and toe of frog is a reminder that an update can move percentages dramatically with no change in the person.
Why Austrian results look odd
- Austria is not a reference category. Reference regions are usually broader, such as Germanic or Central European, so Austria is spread across them.
- The empire mixed populations. People moved across the lands, so Slavic, Hungarian and Alpine ancestry sit together in many Austrian families.
- Neighbouring populations overlap genetically. The statistical difference between an Alpine and a Bohemian ancestor may be small.
- Small shares are fragile. A trace percentage can disappear in the next update.
A common misreading of country pages
Some companies publish pages showing the most common ethnicities among customers in a country. MyHeritage has a feature of this kind, labelled beta, and its page for Austria states that the percentages represent the portion of MyHeritage DNA users in Austria who have that ethnicity. That is a share of people, not a share of an individual’s DNA. If a category such as Balkan appears high on the list, it means that many Austrian customers have at least some of that ethnicity in their estimate. It does not mean the average Austrian is mostly Balkan. Our short overview on DNA and Austrian ancestry makes the same point.
| What the page shows | What it does not show |
|---|---|
| The share of users in Austria with a given ethnicity in their estimate | How much of any one person is that ethnicity |
| Categories defined by the company’s own reference panel | Historical peoples or modern nationalities |
| A snapshot of that company’s customers | The whole population of Austria |
Why siblings can differ
Each person inherits a different, random half of each parent’s DNA. Full siblings therefore share about half of their DNA on average but not the same half, and their estimates can differ in the smaller categories. A brother may show a trace region that a sister does not, even though both descend from the same four grandparents. This is normal, and it is a reason to test more than one sibling if the question is important. It also means that an estimate is a description of your particular inheritance, not of your family as a whole.
Reference panels and updates
A company builds its estimate from reference samples, often people whose ancestors are documented as living in a region for generations. When it adds more reference people or changes the statistical method, the boundaries move. That is why a company can issue an update and shift a customer’s result without any new sample. The size of the Austrian panel is rarely published, and borders in the Habsburg lands were fluid, so Austria itself is not usually a separate category. Read each update’s notes if a company publishes them.
How to read your own result
- Look at the big categories first. Broad regions are more stable than small ones.
- Treat traces with suspicion. Shares under a few percent are often noise.
- Compare more than one company if you can. Agreement across companies is more meaningful than one report.
- Check the date of the update. A method change can explain a sudden shift.
- Compare with the paper trail. If records show ancestors from Bohemia, a Slavic share is no surprise.
Invented example: Imagine an invented woman, Greta, whose four grandparents were all born in Lower Austria according to the records. Her estimate shows 55 percent Germanic, 20 percent East European, 15 percent Balkan and 10 percent other. She is tempted to think the records are wrong. Instead she checks the surnames and finds Czech-derived names on two lines, and a Vienna residence card shows a great-grandfather born in Moravia. The estimate reflects the empire’s mixture, and the records explain it. This is a made-up case.
What not to conclude
- That a small percentage proves a particular ancestor or ethnic group.
- That an estimate can identify a village or a province.
- That a different result at another company means one of them is wrong.
- That an estimate has any bearing on citizenship or legal status.
On the last point, our citizenship guide explains that documents and the law decide.
Using estimates alongside records
Estimates are a prompt for research. If a Slavic share surprises you, the next step is to look for Czech, Slovak or Slovene ancestors in the records, using the guide to Slavic and Hungarian names and the Bohemian and Moravian ancestors guide. If a Jewish ancestry marker appears, our post on Jewish surnames and non-Catholic records guide help. If the estimate is a pure curiosity, the pillar on choosing a test explains how the products differ.
Where to go next
Read what DNA can and cannot show for the wider limits, and how to use matches for the more reliable side of testing. MyHeritage and other platforms are compared on the genealogy records hub.
Estimates change with updates, so a result is a snapshot. Revisit your report when a company announces a new version.
