Evaluation of Evidence
Kory L. Meyerink, MLS, AG, FUGA
In the process of genealogical research, nothing is more important than the accurate evaluation of the evidence found during the research process. Genealogists often use dozens of sources as they seek to solve their research problems and establish proof of a relationship. However, not every source or piece of evidence carries the same weight. Some information found during research is simply more likely to be true than other information. The problem many researchers face, especially beginners, is identifying and understanding the many aspects of each record that must be evaluated.
Unfortunately, the terminology often used to describe a record's reliability is so general that many of these aspects are lost, greatly affecting the way we judge the record and the information it supplies. By more carefully describing the different aspects of evaluation, we can better teach the principles of evaluation. This in turn will lead to better proof of genealogical relationships.
Many instructors in the past have insisted that we class sources (ie: records) as either primary or secondary. But these are not good terms to use when categorizing records and it is often difficult to decide what is a primary source and what is a secondary source. There are two key issues here: the nature of the records themselves (how they were created) and the origin of the information in the record. Simply calling a record primary or secondary avoids the question of how good the information in that record is. A death certificate serves as an excellent example. Most death certificates were created at or near the time of the event, and that definition often is associated with the word primary. However, much of the information on that certificate pertains to events that happened years before the event of death, such as the birth date of the deceased and the parents' names, making the record secondary in regard to these facts.
Because of the difficulty many researchers have in comprehending differences between primary and secondary sources, some have suggested that we do away with artificial descriptions of records that serve no real purpose in evaluating genealogical facts or conclusions.1 There is however, some usefulness in dividing the world of genealogical sources into categories and using descriptive terms for those categories. Using appropriate terminology allows us to discuss the various aspects of evaluation and allows teachers and writers to instruct new genealogists on how they should weigh the various aspects of their findings. There are at least six separate aspects of each source to consider when evaluating our findings: the information, the record itself; the record=s format; and the evidence, facts and events given in the record.
Origin of Information
It is the information in a record that should be considered as primary or secondary. Most records (like the death record noted above) contain both primary and secondary information. For example, the 1880 census identifies relationships at the time of the census and that can be considered primary information. However, the age of adults in the same census is only secondary information, as is their birthplace. Primary information therefore is information that was recorded at or near the time the event (or situation) happened. Secondary information is information that was recorded much later than the event it describes, and/or by someone not closely associated with the event (ie: not in a position to know first hand).
Nature of Records
The nature of the record does not matter to the information. In fact, a compilation of marriages, taken from the original county records, still presents primary information. However, the nature of the source allows additional errors to occur. Here is where a consideration of the categories and formats of records is useful in the evaluation process.
Genealogical records (those which provide information about individuals, such as names, dates or relationships) comprise two categories. Original records were created to record certain events. They were generally written close in time to the events they record, and are usually the earliest record of that event. Thus a baptismal record is an original record, as are obituaries, military pension papers, business account books, and even city directories and most newspaper articles. In each case, the information was written directly into the record by someone who was in a position to have accurate (often first hand) knowledge of the information being recorded.
Compiled records (sometimes called records of previous research) represent a gathering of information from one or more additional sources (either original, other compiled records, or both). The important consideration here is that someone interpreted the information found in other sources. Examples include family histories, biographies, and local histories. Often they are published books, but that is not a requirement. (Neither should all published books be considered compiled records.)
Incidentally, genealogists use other records in their research which do not directly add names to their files. We often forget to talk about records such as gazetteers, dictionaries and how-to books. They are part of another set of records called Reference Tools. Such records also conveniently divide into two categories: Background Information and Finding Aids.
Kinds of Formats
The format of the records also makes a significant difference when evaluating the evidence found therein. Recording errors can appear in any source, but the nature of the errors changes with the format. The actual document (either an original or a compiled record) is often available to the researcher. Virtually as good as the actual document is a photographic copy, including microfilm, microfiche, or photocopy. In either of these cases, the researcher is viewing the record, as it was first made and any errors are the fault of the person who made the record. Copy error may be introduced if the document was transcribed, extracted, or abstracted--an important consideration in evaluation. If such copies are a printed copy the genealogist must account for possible copyist and typographical errors. In a manuscript copy, there may still be copyist errors, but there is now the possibility of misreading the handwriting. In brief, the further removed the copy is from the actual document, the more errors are likely to have been included.
Directness of Evidence
Evidence is the statement(s) of fact made in a record, or the interpretation of the facts in a record. A direct statement specifically states a fact, such as the date of death. An indirect statement (often called circumstantial evidence) reasonably implies a fact.2 For example, a marriage record is direct evidence that a person was born, but only indirect evidence of the time of birth as the person may have been born 16 or 60 years before the marriage. Genealogists usually prefer direct evidence, because indirect evidence may be interpreted differently by others.
The first four aspects include judgments that can be made by examining the single record (or source) alone. The last two aspects include the necessity of judging the information within the record 1) against itself (internal evaluation) and 2) against information from other sources (external evaluation).
Consistency of Facts
Every record includes several facts (or alleged facts). The researcher must judge these facts in conjunction with other facts in the same record or in other records. Consistency or lack thereof determine whether the facts agree with known (ie: proved, believed or accepted) facts, or if they disagree with the other facts. For example, if a birth record of a child shows that it was born a year after the known death of the father listed on the birth record, there is a serious disagreement that must be resolved before that record can be accepted as proof of the child's paternity.
Likelihood of Events
Original records generally document specific events in the life of a person. Compiled records also list events, among other facts about individuals. Part of evaluation requires that the researcher consider whether the events, as shown in the record(s), really may have happened. Some events are less likely than others, such as joining the military at the age of ten or twelve, being born on the father's birthday, or having a probate inventory show a considerably larger estate for a person than recent tax lists or census records indicated. Such events are possible, but unlikely. There is continuum of probability ranging from very likely, to highly unreasonable, to impossible for most situations. This continuum will change with different cultures and time periods.
Evaluation and Proof
Each record can be evaluated individually, based on these criteria, as separate clues, but proof is the accumulation of acceptable clues. The judge (ie: researcher) is responsible for determining if the accumulated evidence represents "clear and convincing" proof of a genealogical fact.3
In evaluating evidence, it is important for all researchers to realize that original records are not inherently better (or worse), than compiled records. However, they do need to be evaluated differently. The same is true with the information in a record. Primary information is not necessarily more correct than secondary information. Every genealogist knows of cases where clerks made mistakes in recording events into original records. Compiled records often correct such errors, making a better record of events and relationships than the originals that they may be based on.
The evaluation must also consider the format, evidence, facts, and events for each record; but each of these aspects must be evaluated differently. Using such descriptors as original or compiled records, primary or secondary information, actual or copied records, direct or indirect evidence, agreeable facts, and likely events becomes a useful way to teach the different ways in which each record must be considered.
1. Paul Drake, "Evidence in Genealogy," Heritage Quest, #40, p. 12.
2. Noel Stevenson, Genealogical Evidence, 1979 rpt. (Laguna Hills, CA: Aegean Park Press, 1989), pp. 181, 186.
3. Paul Drake, "Some Thoughts Concerning Genealogical Evidence and Proof, Part II - Establishing Proof," NGS Newsletter, Vol. 17, No. 6 (1991), p. 153 and Donn Devine, "Do We Really Decide Relationships by a Preponderance of the Evidence?" NGS Newsletter, Vol. 18, No. 5, p. 131.