Pages may contain affiliate links.
See my disclosure page for more details.

Confused by Your Shared Matches?

There are a lot of people out there trying to use DNA to discover their family history and ending up frustrated and confused due to no fault of their own. Are you one of them?





This blog post will cover the first steps if you're dealing with confusing or mixed up shared matches (also called "in common with" or ICW matches).
  • The most common situation making shared matches confusing.
  • How to identify the problem causing your shared matches to be "mixed up."
  • Variations in situations leading to shared matches not making sense.

Family knot id kit
I'm preparing additional posts about the pitfalls of using shared matches (i.e. the problems you cause yourself if you don't i.d. your situation) and the solutions. You need to address the specific situation you're experiencing so it's important to first identify it, then you need to understand the pitfalls so you can recoginze them. Then you can apply a customized fix for the situation and pitfall(s) you're experiencing.

Let's get started!


Your perception: "My shared matches are wrong [or mixed up or just confusing]. They don't match my family tree."
Reality: The root of your DNA/tree problem (ha, ha, root, tree) is you don't have a tree, you have a knot.

It's possible there are other reasons but this is the most common source of confusion and frustration (related to shared matches) aside from flat-out not knowing how to use DNA results.

Don't worry, I'm writing this post because I've discovered lots of very smart people have the same frustration. You're in good company! This issue isn't widely explained (and perhaps it's not well known) that lots of people have this problem and you need to use your autosomal DNA results differently if you have it.

What is a family knot?

I'm using the term "knot" because there isn't a good description out there for this situation.

You may have heard the genetic genealogy terms "endogamy" and "pedigree collapse." Pedigree collapse is meant to be the term to describe this situation but I find it confusing and potentially does not apply to a test taker that is still dealing with this situation.

This post is not about endogamy but it's the same problem. Endogamy is a more severe case. People who know they come from an endogamous population usually know they need to use their DNA results differently.

Just because you don't come from an endogamous population doesn't mean your DNA results will be helpful following standard or "normal" advice.

I know I haven't explained the terms I'm tossing around so if you haven't heard these terms, here's what they mean.

Endogamy is where there is an extreme amount of intermarriage in a closed population (i.e. there are not new people coming into the population to water down the gene pool). This is a situation with an extreme amount of marriages between cousins resulting in having very few ancestors because you have the same ancestors in your tree multiple times. Realize this is an extreme case.

Here's my issue. "Pedigree collapse" is supposed to be the less severe situation. But if you are endogamous, you have pedigree collapse. You have fewer ancestors because they are repeated. So, personally, I avoid using the term "pedigree collapse." Lots of people haven't heard it and it's hard to remember what it means since it goes hand-in-hand with endogamy.

I like using the casual term of a "family knot" because it actually piques most people's interest. It also can apply to the severe situation of endogamy (that's definitely a knot) or a less severe case. The point for me is to get you to discover you need to adjust your approach to using DNA if you have this situation. First, you have to discover you have it, though!

That's enough about terms. What's the problem?

It's Not That Your Shared Matches Are Mixed Up, Your Family Tree is Mixed Up

There's actually a caveat to the statement that your tree is mixed up but let's start with the more straight forward issue.

If you have any "recent" cousin marriages in your family tree ("family tree" in this sense means your pedigree or direct ancestors), your tree is mixed up and you will have confusion using your shared matches. "Recent" can easily be four, five, six, or seven generations back. If you have multiple cousin marriages among your direct ancestors, even though they are farther back, it can still cause a problem.

When I was little there was a silly song called "I'm My Own Grandpa" and that's exactly what this situation is. However, it does not have to be as extreme as the song's situation to cause you issues when using your DNA results. I'm pretty sure that song is about an endogamous population and I want the people who are NOT in such a population to realize their issue.

[Remember: for this post, your tree doesn't matter, it's the test taker's tree. To save time I'll say "you" and "your" for most of this post. If you are like me and you are using someone else's test, make sure and consider their situation, especially if they only share some branches with you.]

As I've started to focus on working on client projects for DNA and then seeking out DNA projects for southern DNA (that is, people from the southeastern U.S.), I've found the scariest issue is being related to matches in more than one way and not realizing it. If you have a known cousin marriage in your tree, you are more careful and less likely to run into pitfalls.

No known cousin marriages? Try the Family Knot i.d. Kit & Workbook to look for branches that are potentially "knotted."


Obviously, you might have a cousin marriage in the next generation beyond what you know but I've found it is very common (at least with southerners) to be confused by your DNA results, mishandle them, and become frustrated, all because matches are related in more than one way and the user never considered this (or if they did, they had no idea they should do something differently or how).

Did you catch that?

It's not just having a cousin marriage in your family tree that causes a problem. This is that caveat I mentioned.

We all have an unknown part of our family tree. With the issue I'm discussing, you will have cousin marriages somewhere in your family tree, it might just be many generations back from your current known tree.

This caveat is, you might not have any cousin marriages in a fairly well researched tree covering a reasonable number of generations (i.e. going back to a point where research becomes very difficult and you will likely not have a complete tree).

Sometimes the research becomes unreasonably difficult because you are at a generation where cousin marriages increase and you can't tell people apart. Maybe that's why you're using DNA.

So realize, this situation, the "family knot" as it relates to DNA, can occur when you see it in the test taker's tree (cousin marriages) and where there are no known cousin marriages. The latter can apply to anyone, not just to someone with little to no known tree.

The most frustrating and undiagnosed issue happens when the confusion is actually due to intermarriages in your matches' trees.

This is why I don't like calling this situation "pedigree collapse." You'd assume that means pedigree collapse in your tree but the more problematic situation is when you can't see anything unusual in your tree.

This is all the same issue but it's how we recognize it (if we can recognize it) that makes such a difference.


The problem is intermarriage between the branches of your (the test taker's) tree. That means marriages outside the expected couples (and of course it includes any time a child is created, regardless if there is a marriage).

If you don't get it, think for a moment.

We're talking about DNA results. You are trying to find new ancestors by identifying people who share the same pieces of DNA with you.

Those pieces of DNA came from one or a few ancestors. The closer you are related to someone, the more pieces of DNA you have and the more shared ancestors.

So, naturally, if we had identified the source of various pieces of DNA, we'd expect this...

Examples


You have a match that shares DNA segments (those pieces) from four great-grandparents. Sounds like a first cousin (they would share the two grandparents, a child from each great-grandparent couple).

Uh oh! #1
Nope, they're a second cousin. They descend from the brother and sister of your grandparents who happened to also marry. No big deal, still a very helpful match. (You might have seen an issue with the amount of shared DNA but since it's all estimates, you can't be sure).

Uh oh! #2 (this is "instead of" #1)
Wait! They share segments from four great-grandparents but not on the one branch. They have multiple relationships to you, none of which is first cousin (at best they'd be half first-cousin). This is where you may be seeing maternal matches who are shared matches with paternal matches and you're confused or thinking your results might be wrong.

Intermarriages Tangling the Branches of Your DNA Tree

Situation #2 is not necessarily something that you will "see" in your family tree. You might have nice distinct branches and it is only where siblings married (that's not your direct ancestors) or a sibling or cousin from one branch married a sibling/cousin/in-law from another branch (also not your direct ancestors).

If this only happened once you might be a bit confused. If this was more recently and there are a lot of descendants who were tested resulting in multiple DNA matches appearing "mixed-up" you might see a pattern of confusion where your shared matches otherwise make sense.

It's when this, marriages between siblings/cousins/in-laws, was common in a population that you have a problem. Just because it's common doesn't mean you're seeing it in your family tree. At some generation, you are likely to have cousin marriages but it can cause a lot of confusion just from the intermarriages in your matches' trees.

The biggest problem is not knowing your tree is knotted, in this sense including collateral lines, not just direct ancestors.


How tangled your knot is makes a difference in how effective standard DNA suggestions are.

Identifying the Problem

If you got to this post because your shared matches are causing you problems, you've probably identified that there is a problem.

If you know there is a cousin marriage in your pedigree, you have this problem.

If you know there are multiple cases of siblings marrying siblings, you have this problem (if that's as far as it goes, it's a minor problem but it likely indicates there are also cousin marriages between your branches, somewhere).

We all have some confusion using the shared matches or in common with (ICW) matches. They are meant to be an easy and quick place to start using your DNA results.

If you have nice distinct branches, it is a good starting place. You might still get some random misleading shared matches but it won't be too bad.

If you feel like all the shared matches are wrong, your shared matches don't make sense, or you constantly find your shared matches misleading, you are probably dealing with a family knot.

Unfortunately, we don't have a list of populations where this is an issue. There are multiple known endogamous populations and if you belong to one, you should seek advice on endogamous populations, specifically.

If your population is interrelated, but not endogamous, following the advice for an endogamous population will make your task more difficult than necessary. Following advice for "normal" populations will also make it difficult but because you're confused, not because the steps are harder.

If you have significant southern ancestry (southeastern U.S.), assume you have this problem. There are several endogamous populations in the south but in general, being southern doesn't cause the severity of endogamy, but it is confusing!

My expertise is southern research but I've seen similar trends when using DNA results for Canadians. I suspect this is the similar migration pattern (early settlers from Europe moved together and had limited introduction of immigrants, unlike what happened in the north, midwest, and west where there's lots of intermarriage with these new groups).

Based only on this pattern, I'd think you might have the same issue in Australia. You would also have the same issue if you're strictly from New England, without intermarriage of new groups.

You can also have this issue in smaller groups.

I had a client with Germans in Kansas. Those Germans all intermarried --- a lot. Because it was over a relatively few generations, this isn't the severity of endogamy. I believe they came from different places in Germany (that wasn't the purpose of the project so I'm not sure) so the DNA issues were only caused by the extreme amount of intermarriage after they settled in the U.S.

If you remember my description of endogamy, it essentially applies (which is why I have issues with the terms we use for genetic genealogy). These Germans in Kansas were a closed community that intermarried for generations. I know the main confusion in my tree is a less closed community that intermarried over more generations.

Endogamy requires enough generations of intermarriage to reach a severe problem of knotting. For it to be endogamy, I believe you will have a direct relationship between generations and population size.

The larger the population, the more generations for it to reach the level of extremity of endogamy. Southerners, of the type that intermarried since colonial America, are a huge population. That's really a lot of people to intermarry with. It causes issues with using DNA but it's somewhere between "normal" and endogamous.

In the example of the (non-endogamous) Germans in Kansas, they had a lot of intermarriage in a pretty small community, but it was only over three of four generations (then people started moving in and out within the U.S.). It's a huge mess to deal with in the DNA but just coming from this population doesn't mean you're dealing with the extremity of endogamy.

This brings me to my final point in this post...

Is Dealing with a Family Knot the Same For Everyone?

No!

There is a spectrum of inter-relatedness between "normal" (each branch is distinct from the others) and endogamy.

Not only can the test taker's situation fall on different points, the project you are working on can fall on different points on that spectrum.

I use my great-aunt's test and her paternal branches are far more tangled than her maternal branches. But, I've been working on a group of shared matches I've determined comes farther back in her maternal tree and it's pretty messy compared to the other maternal branches.

You will be more successful as you start to focus, both by using methodologies (instead of just doing whatever strikes your fancy) and having a goal or specific problem you are focusing on.

You will need to adjust your strategy depending on the level of interrelatedness for the problem you are working on. You may only have issues in one branch but recognizing this and adapting will prevent you from wanting to beat your head against the wall.

There are two related posts for this one. The next post is now available, the other is still in progress. The first shows the pitfalls of using shared matches, specifically with a knotted tree. If you don't quite understand how shared matches work or why you are having problems, that post will help. You need to understand the pitfalls to avoid them so this is a really important topic.

The other post is what you really want (but you have to know what you're dealing with, first!). It covers my suggestions for dealing with the pitfalls of shared matches when you  have a knotted tree.

Grab the free family knot identification kit to help you determine if you have a family knot and where in your tree it is.

If you have a "knotted" by not endogamous population in your family tree, share it in the comments. Let's help others by identifying other groups where confusion in the shared matches is really due to interrelated lines.
 
 Follow-up post: Pitfalls Using Shared Matches or ICW Lists.
 

2 comments

  1. I just ran across your webinar at Legacy and will watch it I have been at the DNA since 2014 when I tested myself and my grandmother and clearly everyone from Barnwell County, South Carolina are related in multiple ways which is why I can't even figure out 2nd cousins to my grandmother!!! I have tested well over 20 relatives. I will spend some time with your information hope it helps me! Thank you

    ReplyDelete
  2. Glad you found this, Kim! I'm really excited the MyHeritage is making the webinar available for free so people can watch and rewatch it as it covers a lot.

    ReplyDelete

Resource Library Links

If a link you click to sign-up for the Resource Library does not work, try this link instead.