Adequately documenting our process matters--it matters to our research and our conclusions. Because it is important, I do get frustrated when sites have search interfaces that are confusing or have interfaces that emphasize "fuzzy" searches that are difficult for the typical research to understand. When one does not understand how a "fuzzy" search works, it makes troubleshooting difficult.
Knowing how a site searches and knowing that a site searches "correctly" is always important for the genealogist. This is especially when research conclusions are based, at least in part, on searches performed on that site. There are times when knowing how searches were conducted impacts the amount of credence we give to a research argument. Documenting that search process is an integral part of analysis. I know there are some researchers who think that documenting the "why" and "how" is unnecessary. I am not one of those. I don't want to know where you parked your car at a library, but do I want to know what name variants you looked for in the records that were there, and how you queried the library's databases in an attempt to find items in various collections.
We will summarize an example.
William Sargent married Ellen Butler in October of 1870 in Davis County, Iowa. There are no later records on Ellen that provide any indication as to her family of origin. Her 1880 census enumeration is her last record period. This leaves only two records providing any information on Ellen--her October of 1870 marriage and her 1880 census enumeration.
An indirect argument as to the family of origin for Ellen centered on the belief that she would have been living in or near Davis County, Iowa, at the time of the 1870 census--conducted a few months before her marriage. While she could have lived elsewhere at the time, it was decided that her 1870 census enumeration location was near her marriage location.
A search of Ancestry.com's 1870 census index was conducted for individuals named Ellen Butler within a one county radius of Davis. A detailed discussion of the search will not be included in this post, but the determination of a possible family of origin for Ellen hinged upon being able to search the 1870 census at Ancestry.com for individuals who matched certain know criteria, specifically:
- name of Ell* But*
- born between 1851-1857
- born in Missouri or Iowa
- probable parents born in Michigan or New York.
- Living in Union County, Iowa, or an adjacent county, in 1870
There are of course potential problems with searching based upon clues obtained in one census. Let's put those aside for just a minute.
Matches to those search criteria were analyzed--all but one of the 1870 hits could be eliminated as being the Ellen of interest. The remaining Ellen was tentatively identified as being the Ellen of interest.
In my "proof" it's not sufficient to say that I "found" a match and that's the one. The citation of the 1870 census entry is necessary, but all it does is help me find that entry. It does not tell me why I believe that entry if "my" Ellen.
Just because one person matched does not mean I have found the correct person. My analysis needs to discuss how I searched the census (including my search parameters, wildcards, etc.), what site I used, and how I eliminated the other matches. All of that needs to be a part of my argument. The citation is a minor part--necessary, but minor.
There is no way another researcher can agree with my conclusion if they are not privy with the website I searched and how I searched on that site. If my search terms were inadequate or wildcards were not included, then my conclusion is not valid as I may very well have not discovered all "reasonable" matches.
