CORPUS LINGUISTIC ANALYSIS OF WORD FREQUENCIES 1 Can Differences in Word Frequency Explain Why Narrative Fiction is a Better Predictor of Verbal Ability than Nonfiction? A Corpus Linguistic Analysis

A Corpus Linguistic Analysis The observation that individuals who read more have better verbal abilities is among one of the most robust findings in reading research. For example, a meta-analysis of 99 studies revealed moderate to strong correlations between reading, measured as lifetime print-exposure several different indices of linguistic competence (Mol & Bus, 2011). (As with all correlational data, causal inferences cannot be supported and alternative third-variable explanations may exist.) These associations were present across a wide range of periods in the lifespan, from preschoolers and kindergartners, through children in grades 1 to 12, up to and including college-and university-aged students (N = 7,669; Mol & Bus, 2011). Notably, print exposure remains a strong predictor of linguistic ability even after controlling for related factors such as age, education, reading comprehension, and 1 Given the robust association between reading and language abilities, it is important to consider more nuanced questions, such as whether unique genres of text might differentially predict verbal ability. To this end, there is growing evidence that literary skills are more closely associated with exposure to narrative fiction compared to Manuscript CORPUS LINGUISTIC ANALYSIS OF WORD FREQUENCIES 2 expository nonfiction. For example, in a study of sixth grade children, self-reported reading of fiction books was found to predict word identification and reading comprehension, but no such association was observed for nonfiction books, magazines, newspapers, or comics (Spear-Swerling, Brucker, & Alfano, 2010). In a separate study on 11 to 16 year-old students, time spent reading fiction predicted a variety of reading-based fiction in predicting verbal ability has also been demonstrated in adults. Mar and Rain (2015) conducted a series of three studies in which they assessed verbal ability using three different subtests of the verbal section of the SAT (formerly the Scholastic Aptitude Test): analogies, sentence completion, and reading comprehension. They found that exposure to fiction—as assessed either by self-reported reading habits or by a recognition-based print-exposure measure—was a consistent predictor of verbal ability whereas exposure to nonfiction was not. Moreover, fiction demonstrated an association even after controlling for demographic variables and exposure to nonfiction texts. A fourth study on synonyms found similar results, demonstrating that this advantage for fiction over nonfiction is robust and replicable (total N across all 4 studies = 960). In light of the fact that these studies examined diverse age groups with relatively large samples, and employed various broad and indirect measures of reading …