Archive for the ‘Genealogical Proof Standard’ Category

“Reasonably exhaustive” research as a process of elimination

One of the most difficult concepts to grasp, when we strive to meet the Genealogical Proof Standard, is the “reasonably exhaustive search.” This requirement demands that we conduct thorough research, scouring every possible source of relevant information touching upon our research question.

The following steps present one way to think about this requirement, as a process of reduction, beginning with a large pool of possible sources and systematically eliminating those that cannot hold relevant information.

(1) Imagine every possible source for information (meaning every written record or artifact that holds information from any location or time period throughout history). This initial pool contains billions (or more) possible records.

(2) Eliminate records that could not possibly hold information about the subject–for example, a record from ancient Egypt or medieval France could not possibly hold information relevant to the life of an African American born in the late nineteenth century in Charleston, South Carolina.

(3) Eliminate records that are unlikely to hold information about the subject–for example, records in states that neither he nor any of his family members (including hypothetical ancestors or descendants), business associates, neighbors, ever visited.

(4) Identify what records remain–what records ever existed in the times and places in which the subject lived–which of these records are currently accessible, and how and where.

So far, this process of elimination still leaves a large pool of possible sources. Examining all of them would clearly constitute exhaustive research. Exhaustive research, however, is not the standard. The standard is to conduct thorough or reasonably exhaustive research. Another way to phrase this standard would be to substitute the term “reasonably” with “rationally.” In other words, we should think rationally about our research question, our research subject, and the historical context of the problem (time and place).

(5) Systematically examine every record within this pool of potential sources. Start with the most common records from the remaining pool–census records, vital records, land records, probate records, etc. As more information about the subject becomes available you may have to add or subtract records from the pool. Be liberal in adding and conservative in subtracting. For example, a deed from 200 years later could provide a title history that includes an originally unrecorded transaction. If three census records provide three separate states of birth for the research subject, expand the pool of potential sources to include all three states.

Again, we must use our ability to think rationally and reasonably to build a case. Several points should bear directly on our use of sources:

  • Be sure to examine any record suggested by information in any examined record. So if a census record states that someone owned land or served in the military, check the land records and military records.
  • Give priority to original records and records holding primary information. This is not to say that we should ignore derivative records or authored works, nor that we should ignore secondary information. During the process of analysis and correlation of the collected evidence we will determine which information appears most reliable.
  • Ease of access is not the deciding factor when considering what sources we use. Records that do not appear online may be available on microfilm that can be rented from the Family History Library or borrowed through interlibrary loan from another repository. Original records that have not been microfilmed may be obtained through the mail directly from the repository or through the use of a local hired agent.

Finally, how do we know when we have completed the process? Is it when every single item on the list has been crossed off?

Sometimes. Again, we have to use our reason and rational thought. If we are trying to identify the father of John Smith and we have conclusive evidence that his father was James, then we can stop the search. For example, suppose we have found the following records through our initial research:

  • A death certificate identifying John Smith’s father as James Smith–the informant being John’s eldest son William;
  • A marriage record, informed by John Smith himself, identifying John’s father as James Smith;
  • Three federal census records, in which John was aged 2 years, 12 years, and 22 years, respectively, living in households headed by James Smith;
  • A will written by James Smith in which he identifies his son John Smith;
  • A church baptismal record that provides John Smith’s date of birth and identifies his father as James Smith;
  • A tombstone, containing the date of birth as presented in the baptismal record and the date of death as presented in the death certificate, thus connecting the two records as pertaining to the same individual.

It is probably safe, in this scenario, to stop searching. We are unlikely to locate further information that will contradict the conclusion that John Smith was the son of James Smith. We do not have 100% certainty of course. A court record or DNA testing could reveal that John Smith had been adopted. Perhaps it is reasonably (and therefore necessary) to search these additional records, depending on the collected evidence in the case. On the other hand, it may exceed the standard, depending on the collected evidence, and therefore be unnecessary. Some of the other relevant evidence that should be considered in this case are the norms of the location, family oral tradition (if any), the socioeconomic status of the family, the birth order of John among the other children of James and his wife, etc.

The successful genealogist must have a developed sense of reason. Most problems that we encounter will not be as clear-cut and simple as the case of John Smith son of James Smith. Far more often we will encounter families where these relatively reliable sources of direct evidence were never created or do not survive. These cases require thoughtful and insightful use of vast amounts of indirect evidence. For these more complex cases, we will have to search many more sources to meet the standard of reasonably exhaustive. We often have to collect and correlate material relevant to every aspect of our subject’s life as well as investigating the lives (and records) of known family members and associates. We must be able to recognize when we have examined enough evidence to make a reasoned conclusion possible.

Genealogical fallacies in logic

When I was in high school I took a course called “Theory of Knowledge.” I may have enjoyed this course more than any other course I had ever taken up to that point and since (right up until I took Course 4 at the Institute of Genealogy and Historical Research).

One of the course units dealt with classical logic, and logical fallacies. I recently found a handout on “Logical Fallacies” that I received in the course. Sadly I cannot produce a “full and accurate” citation to this handout, as I can’t remember details from nearly twenty years ago.

Much of scholarly genealogical research involves the use of logic–deductive and inductive reasoning. Many of the errors that occur in research similarly result in fallacies of logic. I would like to discuss the most common fallacies, as described on this twenty-year-old high school handout, here.

First the definition of a fallacy:

Fallacies are mistakes in reasoning.

A fallacy is an error in reasoning or in an argument. As such they are not rational in the true sense of the word. Most people who commit fallacies do so out of carelessness or haste. Many people who use fallacies are trying to side-step pure reasoning and manipulate others through psychological techniques. There are hundreds of different fallacies, but a few are listed below:

Not only does this definition address fallacies in a simplified way, it also demonstrates that many of the most common genealogical mistakes are directly caused by fallacies. The small selection of fallacies that I have extracted from the handout for this blog post below further demonstrate this connection between genealogical error and poor reasoning:

1. Hasty generalization – Drawing a conclusion (inference) on insufficient evidence.

2. Dicto simpliciter – Not taking genuine exceptions into account.

. . .

5. Circular reasoning – Going around in a circle – using the premise as the conclusion and vice versa – sometimes called “begging the question.”

6. False cause – supposing that two things that are connected by coincidence somehow cause each other.

. . .

8. Ad ignorantium – drawing a conclusion with no evidence one way or the other.

. . .

11. Appeal to authority – trying to get your position accepted because an authority says so when the authority is irrelevant to the issue.

12. Composition – Assuming what is true of the parts is true of the whole and vice – versa – what is true of the whole is true of the parts.

To demonstrate some of these fallacies, let us examine a few scenarios a genealogist might come across:

1. Hasty Generalization. Susan Smith, age 12 years, appears in the 1860 federal census in the household of Richard Smith, age 36 years. The fallacious conclusion would be that Susan Smith is the daughter of Richard Smith, with no further evidence to support it. The 1860 federal census does not provide family relationships, so any conclusion involving family relationships based on this evidence constitutes a fallacious argument.

2. Dicto simpliciter. Alexander Brown, an African American man, appears in the 1870 federal census in Mississippi. A genealogist falling prey to this fallacy would not even look for Alexander in the 1860 federal census, assuming that he was enslaved until the Civil War. The presence of free people of color in Mississippi (though relatively few) constitute “genuine exceptions.” The possibility that Alexander Brown was free in 1860–perhaps in another locality, migrating as a soldier during the War itself–exists.

8. Ad ignorantium. This fallacy is probably the most rampant in genealogy. Suppose family tradition says that your great-grandfather James Lee was related to Robert E. Lee. No connection is found. Or you great-great-grandmother Beatrice Jones was a “full-blooded American Indian.” Again, no evidence suggests this conclusion. Believing what you want to believe despite a lack of evidence is this fallacy in a nutshell.

11. Appeal to authority. A published family history says that William Mulliken came from  West Virginia. The fallacious conclusion would accept this statement as gospel, without further research. Just because it’s in print–even if the author is a well-respected authority–doesn’t mean its true.

We are all guilty of nearly all of these fallacies at one point or another in our career.

Do you have any other examples that immediately come to mind?

If you would like to cite this post:

Michael Hait, CG, “Genealogical fallacies in logic,”Planting the Seeds: Genealogy as a Profession blog, posted 19 December 2012 (http://michaelhait.wordpress.com : accessed [access date]). [Please also feel free to include a hyperlink to the specific article if you are citing this post in an online forum.]

The most important thing you can ever prove

This blog has spent a lot of time discussing the Genealogical Proof Standard, especially focusing on the use and analysis of multiple documents.

The most important fact that you can ever prove—the basis of all of our research—is the identity of the subject of each record. When you find a marriage record for John Smith, how do you know it is the same John Smith that you are looking for? No matter how unique you believe a name is, you still have to prove that the record actually pertains to the subject of your research.

Several situations make this determination extremely important:

Same name, same place, same time. This is the most common issue of identity. One of my ancestors, Henry Hait, was one of three men named Henry Hait or Henry Hoyt living in Stamford, Connecticut, during his lifetime. Learning about my Henry involved being sure that the Henry Hait in each individual record I found was mine, and not some other one. This dilemma is so rampant that it is even commonly called the “same name” trap.

Same name, same age, same place, different times. I believe that this situation may have resulted in more inaccurate family trees than even the more well-known trap mentioned above. Imagine that a man named George Burgess (fictional name) shows up in records from his birth in 1830 through early adulthood in 1850, all in the context of his immediate family. Then another man named George Burgess born around 1830 shows up in records in the same jurisdiction from his marriage in 1851 through his death in 1900. The two men never show up in the same records at the same time. If you don’t take the time to thoroughly examine all available evidence, it is easy to conclude that all of the records pertain to the same George Burgess. Never mind that the reality is that the younger George Burgess—with the rest of the family—moved out west, the year before the older George Burgess moved into town from the next county over. Regardless of which George Burgess you are researching, falling prey to this trap would create an inaccurate family history.

Same name, different place, different time. I call this the “Shaky Leaf” problem. No serious genealogist should ever fall for this one. Just because a shaky leaf (or any other “intelligent” automatic search feature) tells you that a family tree has someone with the same name, doesn’t mean that you should attach it to the tree. (But if you are reading this blog, you probably already know this.)

How do you establish identity?

Each record should contain both internal and external clues as to the identity of the subject. To prove identity, you can use these clues with clues in other records to form a solid body of evidence.

Internal clues are those attributes about the subject that tell you more about their lives: their age, wife’s name and age, children’s names and ages, occupation, birthplace, religious affiliation, etc.

In a recent case, for example, I was searching for a man named Joseph Adams. I was able to reconstruct his immediate family based on later records of his children in adulthood, but had no information about him. Using the ages of his children, I was able to identify his 1830 federal census record among the records of other men of the same name.

External clues are those attributes about the world around the subject that tell you more about them: their neighbors and neighborhood.

Once I had identified Joseph Adams in the 1830 federal census, I proceeded to look for all of his neighbors in the 1830 city directory. Their addresses pointed me to the precise neighborhood that Joseph lived in. Using a map, this allowed me to identify the correct Joseph in that same city directory, which provided his street address and occupation. This then allowed me to locate the correct Joseph in earlier and later city directories. These facts about Joseph’s life begin to form a portrait of the man that will help in resolving identity issues in other records.

Another case study that I am currently revising for possible future publication in the National Genealogical Society Quarterly involves the second problem described above. Without giving away too much detail, I can describe the problem here.

  • A man (“TB2″) lived in Baltimore County by 1783 and died there in 1836. Federal census records prior to his death estimate a date of birth between 1740 and 1750.
  • Another man (“TB1″) lived in Baltimore County by the 1740s, and recorded the births of his children in one of the three parishes that covered then-Baltimore County—including a son named “TB2″ born in 1747.

The full case study not only disproves that TB2 was the son of TB1, but uses other attributes of TB2′s life to identify TB2′s parents and grandparents. No direct evidence connects TB2 to his other family members. It is only by using the internal and external clues in all of the records of both TB1 and TB2 that the true relationships are revealed.

If you would like to cite this post:

Michael Hait, CG, “The most important thing you can ever prove,”Planting the Seeds: Genealogy as a Profession blog, posted 18 December 2012 (http://michaelhait.wordpress.com : accessed [access date]). [Please also feel free to include a hyperlink to the specific article if you are citing this post in an online forum.]

Is primary information truly reliable for genealogists?

“Just how accurate are the memories that we know are true?” “All our memories are reconstructed memories.”

– Scott Fraser, in TEDtalksDirector, “Scott Fraser: The problem with eyewitness testimony,” online video, uploaded 10 September 2012, Youtube (http://www.youtube.com : accessed 12 September 2012).

As genealogists we often look for ways to categorize the records we are using. We call sources original or derivative, based on the generation of the format we are using. We call information primary or secondary, based on the involvement of the informant in the events being reported. These designations are arguably an important aspect of our analysis of facts and details that appear in the records we find about our ancestors.

But how important are these designations, really? How much do they affect a record’s accuracy?

Is primary information necessarily more reliable that secondary information?

Our first sense would be “of course.” Someone who was an eyewitness to an event would be a more reliable source that someone who did not witness the event. This may be true, but just how much confidence can we have in primary information? Can we consider the testimony of an eyewitness inherently reliable?

This is where we must be more careful in our analysis of records. As Scott Fraser explores in the video below, the memories of even eyewitnesses can be flawed in surprising ways. This is no less true in our family history research than in a murder trial. “The brain abhors a vacuum,” Fraser remarks. “The brain fills in information that was not there, not originally stored—from inference, from speculation, from sources of information that came to you as the observer after the observation. … It’s called reconstructive memory.”

Making the designation between primary information and secondary information is a useful exercise in our process of records analysis. It is important to consider the involvement of the informant in the event being reported. However, it is equally important to consider other factors, for example:

  • how much time had passed between the event and the creation of the record;
  • what was the mental condition of the informant at the time of the event;
  • what was the mental condition of the informant at the time of the creation of the record;
  • what potential biases may have affected the reporting of the event, either intentional or unintentional.

As genealogists, finding records may seem like the bulk of what we do. Yet I consider the skilled and knowledgeable analysis of these records to be just as important, if not more so. Finding a record is a small part of the process; understanding the record–what it says and what it doesn’t say, its reliability, its significance–is vital for us to achieve reliable and accurate results.

In other words, determining the reliability of even an eyewitness’s testimony is the only way to determine the accuracy of our conclusions. Part of this process is understanding and considering the nature of human memory.

Take a look at this video for a more detailed exploration of the subject:

If you would like to cite this post:

Michael Hait, CG, “Is primary information truly reliable for genealogists?,”Planting the Seeds: Genealogy as a Profession blog, posted 13 September 2012 (http://michaelhait.wordpress.com : accessed [access date]). [Please also feel free to include a hyperlink to the specific article if you are citing this post in an online forum.]

Free webinar tomorrow – “What is a ‘Reasonably Exhaustive Search’?”

There is still time to register for my webinar “What is a ‘Reasonably Exhaustive Search’?” Legacy Family Tree will be hosting this webinar, tomorrow (12 September 2012) at 2pm EDT.

The first requirement of the Genealogical Proof Standard is that we “complete a reasonably exhaustive search for all relevant records” related to our research objective. This presentation discusses what a “reasonably exhaustive search” entals, why this is necessary, and how to conduct a search. A case study explores how failing to identify all relevant records can lead to missing information and forming inaccurate conclusions about your ancestors’ lives.

To register visit http://www.legacyfamilytree.com/webinars.asp. You will receive a confirmation email after you complete the registration process.

Attendance at the webinar will be restricted to the first 1000 to sign in. There are already many more than this number registered, so it is important to sign in at least thirty (30) minutes early. If you are unable to attend the live webinar, the recording will be available to watch free on the Legacy Family Tree website for ten (10) days.

A more permanent copy of the lecture will be available for purchase on DVD directly from Legacy Family Tree here. Feel free to also pre-order the DVD  whether or not you can attend.

I look forward to seeing you there!

For more articles about conducting a “reasonably exhaustive search,” read:

You can also read any of the articles included in the category “Genealogical Proof Standard” on the right, for details about other very important research and analysis skills.

How do I cite a genealogy research conclusion?

I hope that my recent posts defining conclusions and differentiating between simple and complex genealogical conclusions helped researchers to better understand the nature of what we aim for in our research. As scholarly or aspiring professional genealogists we should always attempt to meet the Genealogical Proof Standard, and its condition that we arrive at a “soundly reasoned, coherently written conclusion.”

When we reach a conclusion the question is then: how do we cite them?

Many genealogists are familiar with the terms direct evidence and indirect evidence. Simplified, direct evidence is evidence–information from a document, for example–that states explicitly the solution to a genealogical problem. Indirect evidence is evidence that does not state explicitly the solution to a genealogical problem, but forms part of a larger proof argument.

These terms coincide slightly with the terms simple conclusion and complex conclusion. A simple conclusion–what is generally called a “fact”–can be reached when we have direct evidence that can be assessed as reliable information. For example, the date of death as reported on a death certificate can generally be deemed reliable. On the other hand, the date of birth on the same death certificate, unless provided by one of the parents, would generally be deemed as far less reliable.

In the absence of reliable direct evidence, we as genealogists must “go the extra mile” to meet the Genealogical Proof Standard in forming complex conclusions. This does not necessarily mean that the conclusion has to be difficult. The proof argument concerning Gabriel Diggs’s date of birth is a relatively straightforward argument. By meeting the conditions of the Standard even on straightforward and simple cases of indirect or conflicting evidence, not only do we strengthen the overall reliability and accuracy of our research, but we also develop the analytical skills and thought processes necessary to attack far more difficult cases.

Back to the original question: how do I cite a research conclusion?

In general, simple conclusions are easy to cite. Each established fact only needs cite the single most reliable source of direct evidence. If multiple sources of reliable direct evidence exist, we can also cite the corroborating sources. How we cite these sources in our written family histories, our genealogy databases, and anything else we produce remains the same.

The problem arises when trying to cite a complex conclusion. If, for example, we reach a conclusion that comes from examining multiple records with indirect and/or conflicting evidence, how would we cite this? Take another look at the Diggs example. The conclusion that I reached through examination and analysis of all sources was not stated explicitly in any single record.

In a professional written product, such as a client report or a case study for publication, we would cite the source of each fact separately. And of course, by fact, I mean simple conclusion. Complex conclusions stem from a series of simple conclusions. In many cases, complex conclusions are built from a series of both simple and complex conclusions. We should write a proof argument in which we cite the most empirical sources for each piece of information, and produce a “coherently written conclusion.” The Diggs example uses this technique for citation.

On the other hand, if we are simply entering data into a genealogy database, or producing a detailed, multi-generation family history or genealogy, then we would use a different strategy.

When entering information into a genealogy database–whether you are an evidence-based or conclusion-based software user–you cannot enter a proof argument as a “fact” or “event” (the terms most software programs use) with a single citation. And even if you could, for a complex conclusion, this would not be appropriate.

In a family history or genealogy being compiled for a general audience (i.e. non-genealogist family members), the focus would likely be on the stories of our ancestors’ lives. While one or two proof arguments might be able to be slipped into the work, we would not really want to include all of them.

In both of these situations,  we should include abbreviated proof arguments, citing each of the records we use, in the reference notes. We should also report any conflicting evidence, any indirect evidence, etc.

Most genealogy database programs now include a notes field that would work perfectly for this purpose. (NOTE: I am unsure about the compatibility of these notes with the current GEDCOM standard.) Endnotes can be appropriate for a general-audience family history, as the information would be there for those who are interested, and can easily be skipped by those who are not interested.

If you would like to cite this post: Michael Hait, “How do I cite a genealogy research conclusion?,” Planting the Seeds: Genealogy as a Profession blog, posted 21 March 2012 (http://michaelhait.wordpress.com : accessed [access date]). [Please also feel free to include a hyperlink to the specific article if you are citing this post in an online forum.]

Simple and complex genealogical conclusions

Earlier this week I posted, “What is a conclusion?,” in which I discussed the application of the Genealogical Proof Standard to every research conclusion that is stated as a fact.

In a comment to this post, Martin wrote,

… Does every event need the GPS now? How boring and tedious do we make a genealogical write-up? Most ages do not agree visa-vis the census. Do we really need to GPS every single person, or can we just conclude that ages in censuses vary? Genealogy is proving relationships. I can understand that proving a relationship might incur the GPS, but every fact or event? I think you’ll run off everyone involved with genealogy including me.

My immediate response to Martin’s comment concluded with the question:

At what point is it acceptable to say, “This fact does not need to be accurate, based on research in all relevant records”?

Ultimately, we do need to ask ourselves this question. My goal in researching–and I assume that many of my readers will agree–is to create an accurate account of my family history. My work for clients also reflects this ideal. However, I can also recognize that for some, genealogy has other purposes and priorities.

On the other hand, I think that it is important to recognize what the Genealogical Proof Standard truly entails. You can read the full definition in my post, “The Genealogical Proof Standard: an introduction.” In short, the Genealogical Proof Standard requires that we:

  • Search for all records that could potentially hold relevant information.
  • Cite the sources we use.
  • Evaluate the information.
  • Reconcile conflicting/contradictory information.
  • Form a logical written conclusion.

As I stated in my last post, every statement that we make as factual is a conclusion based on our research. As such, every fact should meet the above conditions.

However, one will note that I did not state that every fact needed to look like a case study in the National Genealogical Society Quarterly or one of the other scholarly journals.

There is a difference between a simple conclusion and a complex conclusion.

These are not terms one will encounter in genealogical literature, as such. However, in my assertion that every fact is a conclusion, it may serve to define these two kinds of conclusions in this sense.

A simple conclusion is a straightforward statement of indisputable fact.

complex conclusion, by contrast, is a conclusion based on a proof summary or a proof argument.

The following table explores various qualities of simple and complex conclusions:

Simple Conclusion

Complex Conclusion

A simple statement of fact A conclusion based on thorough research
Based on a single document

  • Providing direct evidence of the conclusion;
  • Created contemporary to the event being reported;
  • Created specifically for the purpose of reporting the event being reported;
  • Created by a participant or official eyewitness to the event being reported
Based on multiple facts in multiple documents

  • Providing both direct and indirect evidence of the conclusion;
  • Providing information of varied quality;
  • Created for various purposes other than the purpose for which we are using them
 A single citation to a single record will suffice Each document used to form the conclusion must be cited individually
Corroborating information is not necessary to prove the statement (but may exist) Corroborating information must be examined fully to form the conclusion
Conflicting evidence does not exist Conflicting information may exist, and must be reconciled
May form part of a more complex proof argument Comprised of simple conclusions; may form part of an even more complex proof argument
Example: Person A purchased land from Person B, citing the deed that recorded this sale. Example: Person X was the son of Person Y as illustrated by a proof argument, citing several pieces of evidence supporting this conclusion.

Now I am not the type to reinvent the wheel for no reason. Defining these terms serves a clarifying purpose.

The proof summaries, proof arguments, and case studies that we read are complex conclusions comprised of a series of simple conclusions. If we are to trust the conclusion, we must be able to trust the facts being reported as evidence of this conclusion.

We must therefore recognize the difference between a simple conclusion and a complex conclusion. Many of the proof arguments or case studies that I read are actually comprised of a mix of simple conclusions (statements of fact) and complex conclusions (proof summaries or proof arguments in their own right).

Both simple conclusions and complex conclusions can meet the Genealogical Proof Standard. For a simple conclusion, meeting the Genealogical Proof Standard requires no more than conducting full analysis of the single record being used. Some of the statements in that record would qualify as simple conclusions. Other statements in the same record will not qualify, due to the quality of this record as a source for that information.

For example, stating that Person X lived in County Y on a specific date based on his appearance in a federal census record would constitute a simple conclusion. On the other hand, stating that Person X was born in 1847 based on his age as reported in this same federal census record would not constitute a simple conclusion. The census record is not a high-quality source for this information.

As my last post opined, every fact that is reported as a fact must meet the Genealogical Proof Standard. If you report a fact based on high-quality evidence (as described above), then you have met the Standard for this fact. However, if no high-quality direct evidence exists, then a proof summary or proof argument must accompany this conclusion.

Any genealogical conclusion (whether it relates to kinship or some other aspect of an ancestor’s life) based on a series of complex conclusions that do not individually meet the Genealogical Proof Standard cannot itself meet the Standard. Like the old computer adage said, “Garbage in, garbage out.”

If you would like to cite this post: Michael Hait, “Simple and complex genealogical conclusions,” Planting the Seeds: Genealogy as a Profession blog, posted 26 February 2012 (http://michaelhait.wordpress.com : accessed [access date]). [Please also feel free to include a hyperlink to the specific article if you are citing this post in an online forum.]

Follow

Get every new post delivered to your Inbox.

Join 2,882 other followers

%d bloggers like this: