The lack of quality of Public Ancestry.com family tree s is legendary, to the point many family historians consider them nearly useless. They tend to be so poorly sourced that we’ll bet you’ve checked out and found an ancestor’s name and a detailed birth and death date, but the only source is an Ancestry Member Tree, and when you click that tree, it’s another tree-only, and so on. The worst part of these poorly sourced trees is they often become considered “legitimate” sources because they are repeated so often!

So, we decided we’d be a part of the solution and walk through how we wish all Member Trees were sourced. We’re going to first talk about how sources, facts and citations all work together, how we choose to link them and then what it looks like on Ancestry.com.

Before we get started, please understand this one approach, and it’s our approach. We would never be calling out how someone else is approaching tree sourcing as “wrong”, and this approach isn’t necessarily “right”. It’s right for our research, and if every Member Tree we came across was sourced like this we’d be very happy.

Defining the elements of a good citation, fact, source and proof

Let’s start by defining what’s meant when we’re talking about facts, sources, citations, and the notion of proof.

Citations

Understanding citations, and beginning to enforce the standards you settle on, is one of the turning points as family historians evolve into serious hobbyists. We are HUGE fans of Elizabeth Shown Mills’ Evidence Explained (Link) but at it’s most simple a citation must be a breadcrumb that researchers after you can follow to confirm you work. A well written citation should allow anyone to copy the path to review information themselves.

Sources

Sources are straight-forward as well, for the purposes of this discussion. They are the pieces of information that indicate a fact about one your ancestors. Family bibles, Ancestry.com indexes, headstones, interviews with family members, etc. are all examples of sources that yield clues about your relatives.

Facts

Facts and proof are a little trickier. They tend to both confuse, and be ignored, those newer to genealogy. At their most basic, facts are events that have been proven.

Facts at first seem obvious. If my birthdate is May 4th, that seems like a fact. But facts and proof are intertwined. How do you know my birth date is May 4th? Honestly, other than me telling you it’s my birthday, you don’t.

Even in this simple example, that it’s not your duty as the reader to prove my birthday, it’s my duty to prove that date because I’ve made statement that it’s a correct birthday. For me to do that I can attest I’ve celebrated that day my whole life, my Mom told me it was the day, and there are some family members who were there when I was brought home from the hospital. Additionally I have many Aunts and Uncles who remember my mother being pregnant during the time that corresponds with my birth, and I have photographs of her pregnant that were date stamped during that same time, as well as letters and photos (also stamped) after my birth. I also of course have my birth certificate, which was completed and certified near the time of my birth by the attending physician.

But facts get much fuzzier as we look backwards. For our African American ancestors who died in the late 1800’s, we might have only 2 Census ages to show when they were born. Going back further, we might be relying on various Family History collections that are quoting dates that are 8 levels removed from the original source documents, and those documents are long since lost to history. Of course no one is around to provide a statement that they were present at the time of birth, and rarely do we have historical accounts of our ancestors.

Proof

This leads us to consider how we “prove” “facts” for an ancestor who’s long since gone. In many ways there will never be definitive proof for many of the events in our ancestor’s lives. We’ll hear in court dramas that crimes need to be proven “beyond a reasonable doubt”, but for our hobby the Genealogical Proof Standard (GPS Defined) was designed to help guide us on this question. In the strictest definition of proof, the GPS describes 5 elements needed to argue that a fact is a fact. But we aren’t professional genealogists and the need to formally prove every fact isn’t required, however understanding the GPS will still helps guide us on how close we are to a proven fact, and when we have more work to do. For example, if we’ve just taken a few US Census entries and settled on an ancestor’s birthdate as “proven”, it’s likely we haven’t done a “reasonably exhaustive search”, and so we haven’t hit the first element of the formally proving a fact. That might be just fine for some facts, but just the beginning for others, and as hobbyists doing an Ancestry Member Tree that we aren’t required to be hyper-vigilant proving everything. However keeping the Genealogical Proof Standard in-mind should help you understand where on that spectrum you are.

What to do with Facts?

With the definitions out of the way, and with our basic knowledge of how to create Ancestry trees and ancestors in those trees, let’s get started by highlighting the best practices for sourcing our facts.

Each fact should have at least one source attached

The first key to a solid Ancestry tree is attaching a source for every fact attached to each ancestor. If there’s a fact asserted about an ancestor, but there’s no citation pointing to a source that lead us to attach the fact, no one can prove if it’s accurate or not and we increase the risk of passing bad information along. If we all followed this one rule then all Member Trees would have sourced and cited events and we’d all be able to get much further in our research.

Attach all sources to the preferred fact, or list a fact for each source?

How we decide to attach sources to facts plays a big role in our tree building, and there are two schools of thought on how to attach facts to sources online. The most common method is to attach all facts of the same type (say Birth) to the primary fact, even though Sources may not match that fact exactly. For example, our 3xGGF Wesley has facts that say his Birth was: 16 Dec 1837, Abt. 1836, Abt. 1837, Dec 1837, and Abt. 1838. Most people would be to set his primary date of birth as 16 Dec 1837, and attach all sources to that primary.

The other approach, and the one we use, is to attach each fact to each source as they are sourced. In this example we have chosen 16 Dec 1837 as Wesley’s Preferred birth date and we’ve attached the two sources that indicate that date: his obituary and his family bible. The 1850 US Census has Wesley’s birthday listed as “Abt. 1836”, and so we have an Alternate Fact for his Birth of “Abt. 1836”, with the 1850 Census attached as a source. The same for the 1860 Census and “Abt. 1837”. Repeating the process for each source until we have all known Birth dates (5 in Wesley’s case) listed, and each has the proper source attached.

A lot of people avoid this approach because your ancestors will often end up with many events like dates of birth, where there can be only one, and because all the records can clutter an ancestor’s record. Alternatively, if all the records are tied to a single, preferred fact it becomes difficult researching which sources attached to a fact actually assert that fact.

As an example, as we did our research for this article we found our ancestor Ephraim Treadwell, and when we clicked through to review the sources attached to other’s Member Trees to confirm his place of death, we found each tree had a single Death date, with all sources of his death attached to that one date. A deeper dig showed this is how every of his Member Trees was sourced, and looking at his death fact, it indicates at first glance that every source supports that date of death and that his place of death is Fairfield, CT. However, none of the sources support his death location, and only some support his death day. To determine which sources support which facts, we had to review each source individually, and build a list on paper what each listed as a fact. It would have been MUCH easier if they had chosen to link each source to the fact as it was sourced, but by choosing to link all sources to the preferred death fact, we have to dig through each source to determine what those sources actually support.

In the end, we’re big advocates for showing what the sources support, with facts listed as they are asserted, as the best way to get a true picture of the facts that make up your ancestor’s record. Even when those facts aren’t precise or even correct.

Connect all facts, even when they don’t appear correct

We’ve found that only through the complete presentation of all records can you review and identify what facts are likely correct. Because of that, we prefer to present all our facts as they are sourced, and later interpret what’s likely accurate/inaccurate. If you’re editing out “mistakes” as you’re attaching sources, it’s very easy to make the facts fit your current understanding of your ancestor. It will also leave you blind later when you find a record that might support that “mistake” and you’re missing a new path to truth. This is within reason of course, and if you’re sure the ancestor died in 1767 but a records hint says there is a 1810 US Census entry for him, you can pass on that.

Sticking with our Tradewell family examples, Wesley’s father James Bennet Tradewell has records indicating a range of birth dates from 1790-1799. Originally we didn’t have a single solid record that indicated an accurate birth date, but most of what we had clustered around 1796-97. Seeing the range helped when we found a Family Bible entry for James that indicated 11 Aug 1796. Even though the bible entry was completed at some point after 1855 (meaning it was entered long after his birth and should be treated as suspect) we were comfortable accepting it because it fits the previously known range…which we can easily see in his “Timeline” view.

In the end neither choice is officially right or wrong, but we wanted to put this approach out there for your consideration. It can be counter-intuitive attaching facts that we are pretty sure are incorrect, but for us to better understand what’s correct we need to see the full picture so we can best interpret them to find the truth.

Building a good Ancestry.com family tree

The key to creating a good Public tree is this: make sure you have a source for every fact you attach to an ancestor, remembering that Members Trees are NOT sources. Let’s walk through how we can quickly get started with a new tree.

Start with what you know

We should start trees with what we know. If it’s a personal tree, adding parent’s information and what you’ve heard about grandparents is the perfect start. It’s ok at this point for there to be no sources attached to the facts, you’re just trying to get the outlines of your tree fleshed out with the data you know. Just because mom says grandma’s middle name was Marie, that’s fine to enter for now. If we’re building from a record that lists a new ancestor, it’s the same concept, in that we’ll use the information from the record as-is to start.

Let’s use our ancestor Hezekiah Treadwell (1707-1761) as an example for this process. This ancestor is in a “Working/Uncertified” tree of ours, meaning we haven’t proven the facts and there could be guesses in this tree. We first found Hezekiah as we were building out John Treadwell’s tree, with him listed as a child of John and Abigail (Minor) Treadwell in the Ancestry.com record “History and Genealogy of the Families of Old Fairfield, Vol. II Part II”. This is a good example of building a “skeleton” tree that we might build out quickly to get to a particular, known relative or when we’re building out new family connections as far as we can. They are skeleton because we won’t attach every source and do deep research right away, but we’ll go back later to flesh out these ancestors.

Review your “shaky leaf” hints one-by-one, with an eye towards accuracy

These hints are often only 10% of the records held by Ancestry, but if we’re just getting started on a tree it’s the perfect place to attach the most likely facts for an ancestor. But, just because something is listed as a “fact” on Ancestry, it doesn’t mean it’s either a fact or accurate. Take a few moments to understand the source, give it a quick “smell test” and decide what/how you’re going to use the source.

For example, there is a hint for Hezekiah that references “Connecticut, U.S., Church Record Abstracts, 1630-1920″ and it lists his father as “Jno Treadwell” and his residence on 9 Nov 1707 as Bridgeport, Connecticut. The record passes the first smell test: we think Hezekiah’s father was John, the location is about what we’d expect, and the date listed is within both of their lifetimes. However, when we clicked first on the hint, and then image of the original, we saw the record was actually for Hezekiah’s baptism. The church these records are taken from was in Bridgeport, but we can’t assume they were living in that town, and besides the baptism information is valuable and we would have missed it if we hadn’t reviewed the record completely. Don’t just assume what Ancestry is showing is correct.

The same goes for Find a Grave. Hezekiah’s hints list a Find a Grave entry, which was exciting, but when we reviewed it there is nothing that attaches to an actual grave. It literally shows the Burial as “Burial Details Unknown”, and the Memorial is just a paragraph from the history book we’ve already cited. There’s no unique information in this “record”, and so we ignored it.

Filter out records that don’t provide value

Continuing with Hezekiah, the final two hints we’re provided reference the “Geneanet Community Trees Index” and the “American Genealogical-Biographical Index”. As a rule we try and avoid Index records where possible because they are derivatives of some other work, just one more generation removed from the original. In this case we have a couple of additional issues with these sources, First, there is no way to go back to the source for these indexes because they don’t specifically cite their sources. Second, the Geneanet Index is just a summary of the old public user trees from that old site and we’re trying to get away from unsourced public trees! Finally, the facts they list are also duplicated by other sources we’ve already attached so there’s no value to attaching these indexes.

Attaching Public Member Trees

Wait, weren’t we avoiding these trees?? Yes, but while we can’t depend on them for any facts we do want to link with other users who have our same ancestors in their trees, in case they do later attach a record of value and we want Ancestry to notify us.

When we review these Public trees, the first scan should be for any facts they have listed that we don’t and we can see that by the blue checks next to a fact. For Hezekiah, we see differences in his Birth, Marriage and Death records. Birth is easy to ignore because we know from previously attaching that date the record says he was born either in the towns of Fairfield or Stratford, so we used the county they are both in as his location and made a note of the two towns. The Public tree difference is the location of his birth is listed as Stratford, so that makes sense and we’re not accepting that fact, so we’ll leave the box unchecked. Similarly, the Death value is listing the date of the first court testimony regarding his will, and the court record doesn’t list a date of death. Given that it’s highly unlikely anyone would rush to court to swear in a Will on the same day someone died, we won’t be attaching that date to his death.

But the Marriage value is for an entirely new event, 28 years after his known marriage to Mehitable, and we’ll need to review the trees to understand their sources. Cancel the “New Information” screen, open Member Trees again in Hints, and select the ancestor’s record from the Ancestry Member Tree. That showed us the member’s tree, and when we scrolled down to the new Marriage record, we see it has no Sources attached. We’re not going to attach unsourced facts, so we know we can ignore this event.

Go back to the Add New Information to Your Tree screen, and make sure no facts are attached on the left side. Now, when we click “Save to Tree” this Public tree will attach to Hezekiah, but it won’t attach as a Source for anything and we won’t propagating unsourced facts from other members’ trees!

Attach the facts to the sources

Finally we need to ensure all of our new Sources are attached to Facts, and we have to be careful here because we’ve accepted some hints for Hezekiah with the Source attached to him, but the Source won’t automatically be attached to any Fact. Additionally, Facts often won’t all be linked to the source, like when we accepted the “History of…the Families of Old Fairfield” record above it not only didn’t link to any Face, it didn’t create the listed death year at all. Make sure each of the facts listed in the record exist in the ancestor’s timeline, and attach the Source to each Fact right away. Skipping this step is how we end up with unsourced trees!

You’ve got a strong Ancestry.com family tree…what’s next?

When we’re done we have created a new Family Tree for Hezekiah and Mehitable Treadwell, and their children, that is fully sourced and supported by those sources. This entire process took us less than 10 minutes to complete showing it doesn’t take much time to create an accurate record, even when you have 15 hints. Anyone reviewing your tree, including yourself when you come back later, will be able to easily identify all of the facts relating to this ancestor as well as all the records that support those facts.

If this is a direct ancestor you can do more work to build out their information. Start by running a search in Ancestry, and since you have a solid base of facts the search results will much more focused and likely to be an accurate match. Just attach new sources and facts as detailed above, and your tree will continue to be well sourced.

Our example, Hezekiah, is not likely a direct ancestor of ours but we are building his tree out as we hope to catch more DNA matches. Since we don’t need a full picture, and since we have solid Birth/Marriage/Death info and a complete accounting of his children we will leave his record as-is, and we’ll repeat the process for each of their children, then grandchildren, and great grandchildren, etc. (Matching unmatched DNA Hints by Casting a Wide Net)