As a genetic genealogist, I resist the term "Surname Tests" and prefer the more specific "Y-STR Tests". Surnames (i.e., inherited family names) are a relatively recent development and are not so fixed and immutable as often believed.
- Disclosure: I am biased in favor of Y-STR testing. As a surname project administrator, I've seen it help many further their family histories.
The practice of using surnames came to most of Europe after 1000 AD and became universal in England as recently as 1400. Common surnames often have multiple origins (founders);. many who bear the name will not be genetically related to each other within a genealogical time frame.
Another writer mentioned NPE (non-paternal events) which result in a child bearing the name of other than its biological father's birth name. The effects of NPE tend to be cumulative as the name is passed down to succeeding generations. It's estimated that 30% to 40% of people have NPE somewhere in their family trees. Often, the fact of a NPE is found only after a Y-STR test and can present a means of overcoming a genealogical brick wall..
So what are realistic expectations for a Y-STR test? Chances for matching are certainly improved if one recruits a known descendant of one's direct paternal ancestor to test as well. Only a NPE or wrong genealogy will result in a non-match. The downside is that a match merely confirms what one already knew.
For a "blind test" (potluck) one's matching chances much depend upon the makeup of the database to which your results are compared. A bigger database is better than a smaller one. A database with more people who have a reasonable chance of matching you is better than one with fewer. (For example, Englishmen should probably not test with a company specializing in Chinese DNA.) More inclusive is better than more restrictive.
One's chances also depend on the resolution of the test. A minimal test (12 markers) does not, in most instances, yield enough precision to eliminate ambiguities.
The frequency of your surname is a factor. Rare surnames have a much better chance of matching with the same name than do common names. However, if the name is near extinction, you may not have a close match with anyone living.
Aside: If one has a common, multi-origin name, the Y-STR test can be especially helpful in focusing research on one lineage over others with the same name.
Country of residence is presently another factor. It is true that Americans of English heritage represent only a portion of the English gene pool and that acceptance of this methodology, by comparison, has lagged in England. Participation is growing though and there's no reason why it won't eventually achieve similarly as for Americans of many ethnic backgrounds.
Most DNA surname projects for common names will have about half of participants showing Y-STR matches with surnames other than the one of the project. They can not be grouped into a "Green family" because -- no matter who else they match -- they do not match other Greens.
Expectations may not be met immediately; patience and persistence are recommended. The results have a long shelf-life and may result in matches appearing years later. Often, matches depend on the unknown person who would match you getting tested.
Y-STR testing (or DNA testing of any kind) is
not a substitute for traditional, documentary genealogy; it is merely a complement. To expect DNA to come labeled with names, dates, places and other identifiers is to expect more than it can deliver.
