Its not racist at all. With birth records, you can map a person's name as well as their race, and all kinds of random other data.
If Eliyah are almost entirely named to black babies, it's a 'black' name statistically.
If Cody's are almost entirely named to white babies, it's a 'white' name statistically.
Once you've got a baseline of extreme limits of 'white', 'black', 'hyspanic', 'asian', etc.. names, you can perform blind tests on how people react to different names. There's nothing racist about it. Its pretty simple to quantify.
The hard names are those used by all cultures, which is why researchers don't use those names when performing these types of tests.