Unfortunately it is all but useless to the average user who is faced with classifying sources that aren't easily categorized: a digital copy of an original land record published by a state archives, an abstract of a chapter on your great-great grandfather from a local history published in 1882 and reprinted on a Rootsweb message board, a photo of a tombstone, and so on. The definitions are sometimes hard to remember, too.... why isn't that chapter on my 2nd g-grandfather a primary source? It has loads of direct information about him!
Without guidelines it is really hard for a non-academic researcher to assign surety levels in a way that will be commonly understood by other researchers. I just read a brilliant article by Elizabeth Shown Mills defining standards for evaluating and using historical evidence (you can read it here). She identifies three qualities of evidence that establish their credibility:
1) whether the form of the data is original or derivative, 2) whether the specific data is from an informant with first hand (primary) or secondary knowledge of the matter under consideration, and 3) whether that data represents direct or indirect evidence of the date, name, place, event or circumstance we are trying to prove.
Recently I set out to create a database application that would allow me to manage my research, and -- most importantly -- track the quality of the evidence I am gathering. The criteria identified by Dr. Mills can be described as "objective" - that is, the evidence in question either meets or does not meet the definition. This is key to providing guidelines that can be widely used by researchers with different levels of experience. This is also key to my Genealogical Research System. Rather than focus on adding names to a tree, I start with the assumption that sources are the primary points of interest, and accordingly, it is important to track information about the quality of the physical source itself. My database tool allows the researcher to record each piece of evidence contained in a source and note the quality of that evidence relative to research objectives.
I've just updated the field names in the analysis page of my Evidence library to more accurately reflect the terms used by Dr. Mills (so the templates I uploaded to the Bento template exchange are now officially out of date). Specifically, I've added a checkbox field called "Original source," which the user will check if the source qualifies as an original (actual document or digital copy from reputable source) vs. a derivative. The Source table is supplied to remind you of how the source you examined was formatted. The "Direct evidence" checkbox is checked if the data represents a statement of fact that explicitly supports the evidence, as opposed to data that only provides circumstantial evidence. The "Primary Informant" checkbox is checked if the person supplying the information in the source did so from direct knowledge of the event.
By recording objective information about the quality of specific evidence, the researcher can establish a strong foundation for any conclusions to be drawn from that evidence. This tool also allows users to track their research at every stage in the process, from visits to a repository all the way to final proof of an assertion. As a result, any evidence that is subsequently added to the researcher's family tree software will be supported by more objective measurements of value than the program would otherwise allow.
see Elizabeth Shown Mills, “Working with Historical Evidence: Genealogical Principles and Standards,” Evidence: A Special Issue of the National Genealogical Society Quarterly 87 (September 1999): 165–84; digital image at Elizabeth Shown Mills, Historic Pathways (http://www.HistoricPathways.com : accessed 16 May 2012).