Evaluating Genealogical Evidence

For excellent tips on how to evaluate genealogical sources, see a set of genealogy lessons posted in 1998 to the soc.genealogy.methods newsgroup by Richard Pence, now also on the Pence Family web site, or search the Newsgroup on Google.

Along similar lines below are messages posted by Cheryl Singhal to a mailing list and genealogical bulletin board, the first of which contained a chart about what is considered a “primary” or “secondary” source.  See also an essay on Proof — What Is It?

Area:  Roots-L
Date:  24 May 1994
From: Cheryl Singhal
To:     All
Subj:  Evaluating Evidence

The May 1994 issue of The Generator, the news­letter of the St. Mary’s County (MD) Genealogical Society con­tained the following table to assist in evaluating evidence and determin­ing whether you have a “Preponder­ance of the Evidence” case. I see no author cited.
Primary = 3 points
Secondary = 2 points
Circumstantial = 1 point
Process of Elimination = 1 point

5 points needed to prove a theory.

Civil Birth RecordPrimary
Civil Marriage RecordPrimary
Civil Death RecordPrimary for date, place and cause primary or secondary for age, parentage, etc., depend­ing on identity of informant
Bible RecordPrimary, if made contem­porarily [with event]
Tax RecordPrimary
Court RecordPrimary
Church Baptismal RecordPrimary
Church Marriage RecordPrimary
Church Burial RecordPrimary for date & place; secondary for age, parentage, etc.
Naturalization RecordPrimary
Military Service RecordPrimary
Military Pension RecordPrimary
Original Probate documentsPrimary
Record book copies of probateGood Secondary
Original Land DocumentsPrimary
Record Book copies of land docs.Good Secondary
Newspaper Marriage NoticePrimary
Newspaper Death NoticePrimary for date & place
Secondary for other data
Newspaper ObituaryPrimary for date & place
Secondary for other data
Personal recollectionPrimary or secondary
Tombstone inscriptionSecondary
Census RecordsSecondary
Census indicesTertiary
Passenger listSecondary
Passenger list indicesTertiary
IGI  [LDS Church’s Intl. Genealogical Index]Secondary or Tertiary
LDS family group recordsSecondary or Tertiary
Published genealogical bookSecondary or Tertiary
Published periodical articleSecondary or Tertiary
Oral traditionSecondary or Tertiary or worse

It makes some interesting distinctions which I would not have made, but on the whole, it”s certainly a place to start.

[Defunct] Fidonet Area:  National Genealogical Echo
Date:  08 Apr 1997 02:14:20
From:  Cheryl Singhal
To:  Sarah Nunez
Subj:  Evidence Weighting, #2

OK, the first and largest caution I have is:

As I said, All ratings assume the info is being used to prove what it does prove. To wit: Medical records may be used as Primary evidence of illness or the name of the physician. They may not be used as primary evidence of age, residence, parentage, or such. If these items appear in the file in the handwriting of the patient, they are secondary evidence. In the handwriting of the Doctor, tertiary; in the handwriting of the doctor’s secretary or if typewritten, they are -0s.

Likewise a death certificate is Primary for the date and place of death; it is tertiary for birth date, place and parents. A deed is primary for the names of the grantor and grantee and the date of the sale and the description of the property and its price; it is secondary on the first name in the title chain and tertiary for all succeeding names.

The unrated ones are things which need additional explana­tions from you because I don't know for sure what you mean. F’instance:

Family Business Account Books — probably primary if what you want to prove is that the family had a business; secondary for everything else from the owner’s point of view, and tertiary or lower from the patron’s P.o.V. [point of view]

Farm records — what kind of farm records? the daily logs where the farmer keeps track of how much feed he spread or how many gallons of milk he sold at what price? Or, the ownership record, or the annual list of employees, or the weather log, or what? I’ve never lived on a farm so I don’t have the proverbial clue here.

Blessing certificate?  What’s that? If it is similar to a Baptis­mal certificate for an adult (which in most Protestant churches does not include the age of the party), it’s worth­less except as proof of the event’s occurrence. If it is similar to the Infant Baptismal certificate issued by the Catholics, Lutherans, Episcopaleans and some Presbyterian congrega­tions, then it’s secondary evidence of age. If it is very specif­ically the Mormon Blessing, it’s secondary Proof of age be­cause there is a specific day on which these are to occur, I’m told.

The ones under Court-Related Records ... most of these will be primary for name and locale, worthless for identity. I rated the Attorney briefs as Tertiary as an upper limit: this document says what the attorney says he was told. While most attorneys would not lie to the Court, it remains a recitation of what he was told. Some clients do lie to their attorney.

Under Immigration records, again, if each is used to “prove” the fact and nothing more, they are mostly Primary (i.e., the immunization certificate proves he got one; the change of name proves he changed his name legally; the log book proves that the name appears in it). The exception is the Immigrant Clubs; some of these evidently admitted anyone who spoke the relevant language; thus the fact that someone belonged to the German Immigrants proves only that he spoke German like a native; whether he was German, Dutch, Swiss or Austrian is unknown.

In Church Records, there appear a number of items whose worth depends on the denomination. Removal/Admission notations for, say, Quakers or LDS are far more reliable than they would be for Disciples or Baptists. The report of an LDS missionary in the US, vs the report of a Methodist missionary in Darkest Africa are two widely differing documents. Both are primary evidence that the person submitting the report served as a missionary, and at best secondary evidence of his/her activities. As for the Minutes of Meetings — I have been secretary to enough organizations over the years to consider these works of fiction.

Legal Notices such as Court Claims and Name Changes and Traffic Court — again primary as to the fact that the name appears; worthless for other uses.

Personal papers — what kind of personal papers? You mean things like one’s own birth certificate or divorce papers or whatever? They fall into other categories. Or do you mean things like, newsletters from this organization or that? If that’s the meaning, they’re tertiary.

Correspondence — I have in my possession a letter which refers to a happy, healthy, well-off gentleman who walked out to his garage one afternoon and, for no reason at all, killed himself. Correspondence all too often gives the viewpoint of the writer, not the facts.


[This article originally appeared in the ROOTS-L mailing list.]

Subject:  Proof — What is It?
From:     Bill Yates
Date:     17 Oct 1996


A common misconception among genealogists is that “proof” is some kind of an absolute—that its use establishes a certainty which can never be refuted. A statement often made goes like this:  “John Jones is the father of Tom Jones, and I have the proof.” That is a questionable use of the word “proof,” considered by some to be incorrect. The statement uses “proof” as if it were synonymous with “evidence,” when in fact proof is the use of evidence. In his book, The Nature of Proof, Erwin P. Bettinghaus defines proof as, “the process of using evidence to secure belief in an idea or statement.”1 Thus we see that proof is only as good as the evidence used. To say that you have “proved” your lineage to a Revolution­ary ancestor merely means that you have used evidence to establish your belief that the lineage is correct.

Besides evidence, the other key word in the definition of proof is “belief.” Bettinghaus defines belief as, “any sim­ple proposition that an individual can attest to.”2 Perhaps clearer is the following:  Belief is an idea that someone holds to be true or factual. The important point is that a belief is a very subjective thing, a type of individual judgement. There­fore, the strength of a proof might well depend upon who you are able or might be able to convince with your evi­dence. If the leading genealogist in the country accepts a lineage, the proof of the lineage is apt to be sound. Because he is an authority on such matters, his beliefs will carry much weight. Neverthe­less, it must be remembered that even the belief of an accepted authority does not establish certainty. The only thing that is certain is that the lineage is either correct or it isn’t.

How, then, are we to know who our parents are? Per­haps the word “to know” is a little stronger than “to believe.” The dictionary only leads us in circles. To know is to be sure; to be sure is to be confident; to be confident is to be certain or sure. Can you prove who your parents are? That is, can you use evidence to secure belief that John and Mary are your parents? Certainly, you say, and you proceed to list the following:

  1. I’ve always known it and so has everyone else.
  2. My parents have called me “son” or “daughter” all my life.
  3. I lived with John and Mary ever since I can remember.
  4. My birth certificate says they’re my parents.
  5. I even look like my dad.

Such statements are a list of evidence and nothing more. And they should lead to the belief (by anyone who "believes" you are truthful) that John and Mary are indeed your parents.

As insecure as it may make us feel, we must learn to live with uncertainty in our genealogical work. The validity of any pedigree connection between generations is merely a proba­bil­ity. It is a probability because we are not dealing with iron clad logic, but with subjective judgements. The implica­tions are a little startling. For example, the proba­bil­ity that a given indivi­dual is indeed your ancestor decreases with each suc­ceed­ing generation back in time. To see why this is so requires a little probability theory from the field of mathe­mat­ics. Let us assume that the probability that each link in a pedigree is correct is .90 (1.00 is certainty) or 90%. A 90% probability accumulated over ten generations would lower the probability that all connections are correct to only 35%. Another way of saying the same thing is that there is a 65% likelihood that the man who you believe to be your 8th great grandfather really isn’t! (That is, given the assumptions above.)

The actual probabilities are, of course, not known to the present writer and would be complex to determine. They might be above .90 in some instances. With scanty evidence the probability would be lowered to far below .90. There is one consolation, however. The probability that any random indi­vid­ual is your ancestor increases with time. It is almost a cer­tainty that William the Conqueror is your ancestor if you have much English blood. Genealogy becomes the history of a race after a few centuries.

There are many hidden assumptions in genealogical work. These assumptions account for much of the uncer­tainty which arises. One such is the question of legitimacy, a point well made by the late Donald Lines Jacobus in an arti­cle entitled, “Certainty in Genealogy,” which he published some years ago. If any form of inference, such as a logical deduc­tion, is used as proof, the validity of the assumptions as well as the validity of each step in the deduction must be con­sidered. The following example is given so that the reader can look for the hidden assumptions and can better understand the nature of proof in genealogy.

Statement:  John Williams was the father of Elizabeth Williams.

Establishment of Identity:

Elizabeth Williams

John Williams


  1. Elizabeth Williams’ name appears on a list of eight ap­par­ent siblings with birth dates given. (old handwrit­ten record on a sheet apparently torn from a bible) (3)

  2. Bartlett J. Wilson’s family is listed in the 1840 census two lines from John Williams' family. (4)

  3. Deeds show that the estate of John Williams was di­vided into seven parts at his death. Two of the names on the list mentioned in #1 are shown to be heirs of John Williams, namely Lydia and Isaac. (5)

  4. Age groupings in the 1830 and 1840 census of John Williams’ family match the list of #1. (6)

  5. Bartlett J. Wilson married Elizabeth Williams at the home of John Williams. Traditionally, couples were married in the home of the bride’s father. (7)

  6. A granddaughter and contemporary of Elizabeth Williams named some brothers and sisters of Elizabeth which matched the list in #1 as well as one of the names given in #3. (8)

Much of the evidence given in #1-#6 is circumstantial, which, however, is convincing in support of the original belief.

An Inference:
  1. From #3:  Isaac Williams was the son of John Williams

  2. From #6:  Elizabeth Williams was the sister of Isaac Williams.

  3. Conclusion:  Elizabeth was the daughter of John Williams.

The Belief Which Results:  John Williams was the father of Elizabeth Williams.

The forgoing is nothing less than a formal proof that Elizabeth Williams was the daughter of John Williams. The reader might ask, why was it necessary to establish the ident­ity of Elizabeth and John? Identification is the use of facts to establish with high probability the identity or unique­ness of a specific person. Vital statistics (birth, death, marriage) are frequently used because they are commonly available. These facts label a person as unique. Banks ask for name, birthdate and birth place, and mother’s maiden name because it is al­most impossible for two people to have these same statistics attached to them.

Identity of an individual is one of the hidden assump­tions which were formerly discussed, and one which com­monly causes grave errors to occur in genealogical work. The fewer facts known about an individual, the greater the likeli­hood that he is not unique, and that he can be confused with someone else. Thus we see the importance of learning every­thing possi­ble about our ancestors — especially in the ab­sence of vital statistics. Facts should be gleaned from every possible source. Such a procedure increases in impor­tance as the frequency of occurrence of the name increases. In the Southern states such meticulous research often is necessary due to the lack of public vital statistics.

In the proof given above it was assumed that the dates and places listed for identification were indeed correct. With­out realizing it we go through a process of proof anytime we attach a date to someone’s name. Of course, the dates in and of them­selves have no significance. They are important because they often help with identity and therefore help to prove pedi­gree connections. There are assumptions relative to the evi­dence proving Elizabeth to be a daughter of John Williams, first that all the evidence is accurate, and second that is rele­vant. Nevertheless, the proof does exist, and does indeed establish a belief in the mind of the present writer, at least, that the claimed relationship is a true one. Belief in this case leads to action — the writer will continue in his search for the wife and ancestry of John Williams, believing as he does that John is his 3rd great grandfather.

  1. Erwin P. Bettinghaus, The Nature of Proof (Indianapolis and New York:  The Bobbs-Merrill Company, Inc., 1972), p. 8.

  2. Bettinghaus, p. 17.

  3. Photocopy of an original page with names and birth­dates of members of the Williams family. The original was in the possession of Della Turner, Squires, MO and had the appearance of age, with old handwriting and faded ink.

  4. Federal Census of 1840 for Marion Co, TN, microfilmed, p. 4.

  5. Marion Co, TN Deed Books F, p. 446 and G, p. 245.

  6. Federal Census of 1830 for Marion Co, TN, microfilmed, p. 51, and of 1840, p. 4.

  7. Civil War Pension File of Bartlett J. Wilson, WO 427331, National Archives, Washington, D.C.

  8. Zona Sell, Glenwood, IA (now deceased), letter dated 24 Jan 1965. She was born 21 July 1878, and was thus nearly 22 years old when Elizabeth (Williams) Wilson died.

Originally published in The Ridge Runners 4:243 (1976).
Copyright ©1976, 1996 by William A. Yates. All rights reserved.

Bill Yates
Yates Publishing
P.O. Box 67
Stevensville, MT 59870