To the contrary, in a recent interview (maybe on the Jist or Marketplace? I can't recall exactly where I heard this...) the author mentioned that the distinction between "nigga" (a common lyric) and "ni**er" (not a common lyric) made it easier to distinguish potentially racist searches from others. On the flip side, the author ran into trouble when trying to study sexist/misogynistic searches, as many of those are people looking for porn.
It should also be noted that the punchline is not "people who search offensive phrases are racist." The punchline is that seemingly racist searches correlate (i.e. the effect is statistical, rather than individual) with other variables (such as regions where Obama underperformed when compared to other Democrat candidates and/or polling) that seem to indicate some underlying racism.
The actual book appears to be pretty nuanced. The Vox interview linked above is also appears to be relatively nuanced. The Slashdot summary and the paragraphs preceding the interview on Vox are sensationalist, click-baity claptrap.