Booker T. and George: What’s in a Surname?

The media are notoriously reluctant to mention the race of criminals. The pious fiction is that since race doesn’t matter no one is interested in it. On the contrary, everyone, from whatever perspective, is interested.

Fortunately, it is usually easy to find out the race of a criminal. On-line versions of newspapers often include photos, and many names are a giveaway. Laqueesha Jackson can only be black, and Miguel Garcia is bound to be Hispanic.

The first names of blacks and whites have been diverging sharply. Records show that 100 years ago the 20 most popular given names for blacks and whites were virtually the same. That began to change in the 1960s, when blacks started giving their children distinctly black names. Some, like Rasheed or Karim, were supposed to be African, but most, like Shaneequa, Latonya, or DeShawn, are home-grown, black-American innovations.

Innovation marches on. In California in 1970, the typical black baby girl got a name that was twice as common among blacks as whites. By 1980, her name was 20 times more common among blacks, and by 2004 more than 40 percent of black girls born in California got a name that was not given to a single one of the 100,000 white girls born in the state that year.

Everyone has noticed this trend in first names, but did you know that the US Census Bureau keeps track of what last names tell you about race? It has a database of the 1,000 most common surnames that appeared in the 2000 census, along with the percentages of people of different races who had that name. If you go to this link and download the Excel file called “File A: Top 1000 Names” you will get the full database. It is fascinating to sort the names and see which are most likely to be white, black, Hispanic, etc.

To start with whites, there are 146 names out of the 1,000 most-common that are 90 percent or more white. The whitest name of all is Yoder (98.11 percent of the people with that name are white), followed by Krueger (97.06 percent), Mueller (96.96), Koch (96.86), and Novak (96.84). The entire list of at-least-90-percent-white names can be found here.

The second and third columns in the table, “Rank of Name” and “Number with that Name,” indicate how common the name is. In 2000, Yoder was the 707th most common name in the country, and 44,245 people had it. You will see that even among the Yoders, there is a tiny number that are black, Hispanic, API (Asian/Pacific Islander), multi-racial, etc. It would be interesting to know how the 80 Asians (44,245 x 0.0018 = 79.641) who have the name Yoder got it.

As you can see, the whitest names are Scandinavian, Jewish, and Irish. This is probably because most such immigrants came after the abolition of slavery or shortly before. Some slaves took the family names of their masters and many named themselves after famous Presidents. It is safe to say that most former slaves had never heard of the name Yoder.

There are 791 surnames — nearly 80 percent of the most-common1,000 — that are majority white. The last one on that list is “Robinson,” which is 51.34 percent white and 44.1 percent black. My name, Taylor, is the 13th most common in America and is 67.8 percent white and 27.7 percent black.

Which names are most likely to be black? This table lists the top 100. It is well known that Washington is a common black name, but would you have guessed that no fewer than 89.97 percent of the 163,000 people with that name are black? Or that 75 percent of the people named Jefferson are black? Besides these names there are seven more that are majority black: Booker (65.57 percent black), Banks (54.24), Jackson (53.02), Mosley (52.83), Dorsey (51.81), Gaines (50.27), and Rivers (50.21).

There are 105 names that are 90 percent or more Hispanic. Another 55 names are over half Hispanic and then there is a sharp drop off in frequency. The 164th most Hispanic name, Costa, it is only 8.48 percent Hispanic, with the rest mostly white.

Only 22 names are at least 90 percent Asian, and then only six more names are more than 50 percent Asian. Like the Hispanic names, they are all easily recognizable.

It is possible to sort the Census Bureau database to find the names most likely to be held by people who say they are American Indian/Eskimo or multi-racial, but the percentages are so small this information doesn’t tell you much.

Lowery is the name in the top 1,000 that is most likely be Indian or Eskimo, but since 96 percent of the people with that name are not Indian or Eskimo (70 percent are white and 22 percent are black) this does not tell you much. A name like George Running-Bear is likely to be Indian, but Running-Bear is not common enough to be in the top 1,000 names.

The 10 most common names for people who claim to be multi-racial are Ali (17.5 percent of Alis say they are multi-racial), Kahn (15.6), Singh (15.3), Shah (5.9), Patel (5.8), Joseph (5.3), Costa (5.2), Andrade (5.0), Silva (4.8), and Vang (4.8). There seems to be an unaccountably strong Indian/Sikh contingent.

Trying to judge race from last name is purely a matter of statistics, and some names are notoriously ambiguous. The name Lee is 40 percent white, 17 percent black, and 38 percent Asian. Sometimes, though, the statistics are convincing. If you are trying to rent an apartment and you get a message from someone named Mueller or Barajas or Jefferson or Choi, you can be pretty sure with whom you are dealing.