The paper's primary purpose is to outline an experimental logic that (1) considers causally prior skills such as nonverbal IQ, vocabulary, and decoding, and (2) emphasizes the relative importance of component reading skills rather than simple differences between groups of good and poor readers. (Author/DB)