Individuals who read more tend to have stronger verbal skills than those who read less. Interestingly, what you read may make a difference. Past studies have found that reading narrative fiction, but not expository nonfiction, predicts verbal ability. Why this difference exists is not known. Here we investigate one possibility: whether fiction texts contain more of the words typically evaluated by verbal ability measures compared to nonfiction texts. We employed corpus linguistic analyses to compare the frequency with which commonly tested SAT words appeared in both fiction and nonfiction texts, for 3 different corpora. Differences in SAT word frequency between the two genres were found to be negligible across all corpora. As a result, we conclude that there is little evidence that differences in word content between fiction and nonfiction texts can account for their differential relation to verbal ability. Other possible explanations are proposed for future research.