Surname Frequency and YDNA Studies

Return to The Tapestry Homepage		Return to Methods
Contents 1 Documentation 2 Background 3 Data 4 Discussion	This is one of a series of articles on Genealogical Methods, prepared in association with The Tapestry. See Index for a list of related articles.

Documentation

Background

This is an analysis of the frequency with which certain certain appear in the 1810 census, and the number of participants in FTDNA YDNA projects. Ancestry's US Census data were used to estimate the frequency in the 1810 census, of five surnames that have been examined on the Tapestry. Data was extracted for a "Standard Match" of "All" variants of the surname( as picked up by Ancestry's search engine), and for "Exact matches" only. In two cases, the "Houston/Huston" and "Cowan/Cowan" surnames, where there are known common variants in use in 1810, exact matches for both variants were recorded. (Note the 1850 hits for the standard match for the snoddy surname, compared to the 14 hits for the exact match, indicates great variability in the Ancestry "Standard match" algorhythm m for variants of this surname. That suggests that Ancestry's algorhthym for this surname is weak. Direct inspection of this search showed that the results included numerous hits for names (such as "Smith") which would not be normally considered to be variants for "Snoddy".

In addition to surname frequency data were extracted from the FTDNA site with respect to the number of persons they show for the various surname YDNA projects. It is assumed that these data include all reasonable variants of the surname.

Data

Surname Number of Persons 1810 Census Number of Tests Total Hits Exact Hits Cowan 1850 109 151 Cowen 101 Houston 941 125 86 Huston 167 Kilgore 75 40 31 Snoddy 12,356 14 10 Walker 2007 1777 789

Discussion

The above data represents a small sampling of the universe of surname and YDNA data. The Walker surname, for example, is among the most common surnames in the United States, while the others are all relatively rare. This study would benefit from increasing the sample size to include additional surnames at the intermediate and large scale range of surname frequencies. There may also be a need to expand the number of "surname equivalents" (such "Huston" vs "Houston") used for data capture in the 1810 census. Doing so would probably improve the precision of the results; for present purposes the data appear sufficiently consistent to not require this.

These data show a surprizing degree of consistency in terms of the frequency of a given surname in the 1810 population, and the frequency with which descendants in today's population take the FTDNA YDNA test. Thus, uncommon names seem to receive no greater or lesser attention on the part of descendants than very common names. Whether this conclusion would hold-up with more data, particularly for the mid and large scale surname frequencies, is something that might be further explored.