A significant proportion of American high school students struggle with reading comprehension. Theoretical models of reading comprehension might help researchers understand these difficulties, because they can point to variables that make the largest contributions to comprehension. On the basis of an extensive review of the literature, we created a new model of reading comprehension, the direct and inferential mediation (DIME) model. The model hypothesizes relationships among background knowledge, inferences, reading comprehension strategies, vocabulary, and word reading and addresses the direct and mediated effects of these 5 predictors on comprehension. The authors tested the fit of the model and 3 variations of the model to data from 175 students in 9th grade. The DIME model explained 66% of the variance in comprehension. Vocabulary and background knowledge made the largest contributions to comprehension, followed by inference, word reading, and strategies. Analyses of participants scoring below the 30th percentile on comprehension showed these students to have low scores on all of the measures. The authors suggest that vocabulary and background knowledge interventions might be the best way to begin improving the academic reading comprehension of students like those in the sample.