The Influence of Reading on Vocabulary Growth: A Case for a Matthew Effect