Reading Errors of Importance: The Role of Word Frequency and Relevance in Passage-Specific Comprehension

Reading efficiently and accurately is critical to reading comprehension. Determining what words are important for readers to know and be able to read accurately to support passage-specific comprehension is less understood and has been conceptualized differently throughout the literature. Two ways of conceptualizing word importance, word relevance, as measured by Term Frequency-Inverse Document Frequency (TF-IDF), and word frequency, were the focus of the current study. 191 adolescent participants (97 males and 94 females, 9-14years old) read two informational passages and the number of errors was recorded. Each word's frequency and TF-IDF per passage was determined. Random effects logistic regression models were used to determine if word frequency and word relevance metrics independently predicted the number of errors for each passage. To examine how word frequency and relevance impact passage-specific reading comprehension, each error was classified by frequency and relevance level (e.g., low relevance-mid frequency) and then the number of errors in each category was used to predict passage comprehension outcomes. Results revealed that the two concepts of word importance are not redundant in predicting reading errors. The number of errors on low frequency-high relevance words predicted passage-specific comprehension across passages (β = -1.04, p = 0.04; β = -1.28, p = 0.025). Word relevance is a viable metric to conceptualize word importance in texts. Considering word relevance in addition to frequency could have research and practical implications related to passage comprehension.