Evidence-based Genealogy vs. Conclusion-based Genealogy

Recently there have been some interesting discussions among some of the Genealogy blogs I frequent, concerning Evidence-based genealogy and Conclusion-based genealogy, the first being a top-down approach and the other, a bottom-up. Of course, depending on your point of view, you might consider the first bottom-up and the last top-down, the semantics don’t really matter. The distinction between these approaches can be difficult for many of us to grapple with, especially since most of us tend to approach genealogy from both ends.

Evidence-based genealogists (EBGs), like myself, don’t add claims to their database unless there is some evidence to support the claim. Generally this evidence comes in the form of sources. EBGs will add their sources and claims as soon as they become available, all times being relative. This approach will lead to multiple and duplicate claims for individuals even for claims which are considered single-occurrence-events (SOEs), such as a birth date or a death date. Obviously, these claims cannot all be factual, but, since EBGs are not focused on a conclusion model, this is not an obstacle.

Conclusion-based genealogists (CBGs), which make up the bulk, take a different approach. They gather all their sources together over a period of time, and when they feel comfortable with making a conclusion based on the sum of their data, they will enter only those claims for which they have made conclusions, into their database with their supporting sources. Using this approach, the database will usually not contain multiple claims for SOEs.

As I’ve said, I’m an EBG. The reason for this is in the way I use my database. I publish my database online at The Forsythe Saga website. It is my belief that if you are making your data available to a large number of people, you should give them all the tools they need to make their own conclusions. Having been doing online genealogy for over a decade, I have come across thousands of conclusions made online by other genealogists that I’ve subsequently found to be factually incorrect, or they are not documented so that there is no way to verify their conclusions. Unfortunately, without the proper evidence, as a EBG, their unsupported conclusions are worthless to me and must be discarded.

Genealogists who do not publish their data online, again the bulk, are more interested in conclusions. The reason for this is that they usually share their data with other family members who are not so hard-line as me, and are only interested in conclusions. My problem with this approach is that if I discover new information that conflicts with someone else’s data and that someone has already concluded that the claim is not valid, and therefore left it out of their database, I do not know that and must redo all the necessary research to arrive at the conclusion myself, which I don’t mind doing, but I wish I had their reasoning at hand.

Both EBG and CBG models are valid approaches, and most of us intermingle the two. Both types of genealogists are interested in conclusions and will look for conflicting claims in their data, and then often focus their research in those directions (See Bonkers, my new online tool that functions as a GEDCOM sanity checker). They will also both develop proof arguments to resolve conflicts and include these as part of their research.

In this modern era of computers and the Internet, the sharing of genealogical data with strangers becomes more and more prevalent. I encourage anyone who is currently focused on the conclusion model only, to begin thinking about the advantages of an evidence-based approach to entering their genealogy data.

Louis Kessler in his post I Have Evidence About the Future of Behold has suggested a new paradigm for genealogy software that would better support evidence-based data entry by allowing claims to be added directly from a source and then assigned to individuals. His intention is to make this approach available in his Behold application when editing capabilities are added. I commend him for this.

Until that time, a simple set of guidelines can be used for EBG data entry using any genealogy application that allows sources to be added independently of individuals.

1. Enter your source, including all information related to it such as publishing information, repository, and all quotations/citations.

2. Add source categories to your source. . If your genealogy application does not support adding categories to your sources, then you should use the GEDCOM source reference quality settings when adding your claims below.

3. Now that your source is in place, scan through the text of your source and look for any claims it makes. Many people, myself included, will concentrate only on claims pertaining to their particular family or ancestors, nothing wrong with that. Add the claim to the individual or family it applies to and be sure to reference the source and set the source quality if you were not able to categorize your sources. If your application supports adding a page number to the citation, go for it.

4. Repeat step 3 for all claims made by your source that you wish entered in your database.

Your done!

To better understand the benefit from using this approach, look at the claim tab for any of the sources on my website. The claim tab separates the claims into 5 types: names, genders, associations, events and attributes. The claims will be displayed on the tab in that order. Having all of the source claims available and grouped together like this allows me to easily review my source at anytime to determine if I have captured all of a source’s relevant information.

Other blog posts related to evidence/conclusion based genealogy:

Michael Hait, What is a conclusion?

Randy Seaver, Events, Assertions, Evidence, Facts, Sources, Analysis, Conclusions, Software, Oh My!

Russ Worthington When to enter data into your Genealogy Software

Randy Seaver, Are You an Evidence-Based or a Conclusion-Based Genealogist?

Louis Kessler I Have Evidence About the Future of Behold

Randy Seaver Follow Up Friday – More on Evidence-Based and Conclusion-Based Genealogy

Michael Hait Simple and complex genealogical conclusions

Randy Seaver More on Conclusion-Based and Evidence-Based Genealogy

Randy Seaver Trying to be an Evidenced-Based Researcher- Does This Work?

Randy Seaver Does Genealogy Software Force Me to be Conclusion-Based?

Tim Forsythe GREnDL 2.0 and Counting

Louis Kessler Evidence and Conclusion Modelling in Behold

Randy Seaver Puzzling Over the Evidence-Conclusion Process

Ginger Smith Am I an Evidence-Based Genealogist or Conclusion-Based Genealogist?

GeneJ Genealogy Software: The Evidence and Conclusion Process debate

James Tanner Source Citation vs. Proof

Jennifer Zinck Evidence Management

Randy Seaver, Follow-Up Friday – Reader Comments on Evidence and Conclusions

Michael Hait “Evidence-based” and “Conclusion-based” software use

Leave a Comment
  • Tim Forsythe
    15 Aug 2014


    I agree with you that conclusions do not always match the summation of the evidence. That is of course the point. I propose that claims based on evidence and those based on conclusions be explicitly indicated. I would hope that conclusions are always accompanied by clarifying text. Together, reviewers should be able to understand and interpret the data. Since and evidence based claim simply indicate that a source made the claim (rightly or wrongly), and are not making a conclusion as to whether or not the claim is valid, it is generally not disputable (i.e. the source made the claim or it didn't - making it disputable). Conclusions on the other hand are disputable based on the interpretation of the evidence. What we don't want to see in our databases are conclusion based claims without supporting evidence based claims and clarifying text, because they provided no reference for the reviewer.

  • Peter Booth
    15 Aug 2014

    Its a fascinating, crucial debate, though I'm uncertain as to how well the concepts that I see referred to (claims, evidence, conclusions, proof) match the complexity of reality. For example:

    I believe that my 2nd great grandfather, James Catterall Fearns, married a "Mary Effrom" on 20 May 1866 at St Mark the Evangelist Church, Witton (Blackburn) England. The marriage return says that Mary's mother was a tailor Francis Effrom and the groom;s father was a flagger/slater called Henry Fearns. "My" James Catterall Fearns was the illegitimate son of a Catherine Fearnes who married a flagger/slater Henry Catterall. The census returns of 1871, 1881, and 1891 support the idea that they had ten children beginning with Elizabeth Catterall Fearns (1867) and ending with Mary Catterall Fearns (1888). The baptismal records for these children support the idea that their mother's name was Mary Ann and that she was born in Liverpool. The Lancashire birth register shows 9 children born in Blackburn at this time, with the name Fearns, who had mother's with maiden names: Hefferin, Heifer, Hefferan, Hefferin, Holden, Heffran, Heffron, Hefforn, and Heffron respectively.

    There are census returns and marriage records that I believe refer to the tailor Francis Effrom who was Mary Ann's father. These documents refer to Francis Heffarn, Heron, Hefferon, Haffren, Sefton, Haffren.

    In other words every source makes a slightly different claim, and my conclusions, whilst based on the collection of sources, don't directly mirror the claims in the sources. My current "theory" is that "Francis Heffron had a daughter Mary Ann Heffron, who had ten children with my great grandfather." But there is a wrinkle- The 1851 and 1861 census returns show Mary Ann as being born in Manchester, yet the latter four returns 1871 to 1901 say Liverpool. This could be a human error, it could be an intentional misstatement - or it could also mean that the last four returns are treferring to a different person than the first two. This latter alternative would mean a whole different interpretation of the data, and would lead to different conclusions. In this example, I have seen two researchers look at the same data and draw different conclusions.

    I would like my software to show me at what tome I arrived at a conclusion and what eveoidenbce/data was available at that time. If I then collect additional data that seems to refute a conclusion, I'd like that to be highlighted so I could re-examine and possible modify my conclusion. I also like to be able to run with a couple of alternative explanations and see where they lead me. This is a bit like creating branches in git for exploratory work, that often gets thrown away at the end of the day.

    Sometimes, when I'm in "exploratory" mode I will create one-off family trees in ancestry.com and incorporate a bunch of poorly documented public trees and then export to familysearch.org sand see how much of their persona data supports the ancestry tree's conclusions. Often it leads nowhere but sometimes it allows me to rule out whole family lines and get to the "real" picture. Sometimes someone mails me and asks "why did you put this X as Y's father? If you look at the scanned census return you'll see that they forgot to extract the record about the twin brother philip, which is why you didn't see Z..." I will revue this and fix any mistakes that I find. The problem is that those mistakes included some intellectual work (faulty conclusions) that is now gone forever.

  • Tim Forsythe
    11 May 2012

    In general, GEDCOM supports two 'text' fields for sources. TEXT, which is intended for including actual quotations, and NOTE, which is intended for including comments. I would imagine this is how PAF will export your source fields to a GEDCOM file. This is easily enough to check by exporting your file and then examining it with a text editor to see how those two fields are exposed.

    Unfortunately, as I stated in my column, GEDCOM allows any number of NOTE fields, but only a single TEXT field per source. For this reason, I personally use multiple NOTE fields, 1 for each quotation, and the single TEXT field for my comments (things like full title, more publisher info, who contributed the info, etc.). Ancestors Now trees will, of course, show both fields, and is non-restrictive as to the number of TEXT fields you may have (most genealogy apps will limit you though, I don't remember if PAF did or not. It's been a long time since I used it).

  • borgsteede
    11 May 2012

    I'm trying to accomplish something like this with PAF, but I don't see how I can add notes to sources there. When I enter independant sources, i.e. sources that are not linked to events, I only see two fields that seem to support multiple lines, and those are named text and comment. Any idea how I can use those with your site?