Date of Award
Doctor of Philosophy (PhD)
This thesis undertakes an analysis of current theories of fluent reading in Section 1. This analysis leads to the recognition of fluent readers as flexible. They can use both recoding and whole-word single word identification skills as required (LaBerge and Samuels, 1974) as well as use higher-order (top-down) semantic and syntactic knowledge in interaction with word identification (bottom-up) skills (Rumelhart, 1976). In Section 2, theories of why some children have difficulty learning to read are considered with respect to the conclusions drawn in Section I. The tentative conclusion here is that poor readers do not lack the flexibility required to use higher-order skills in contrast with lower-order skills but that poor readers do lack the flexibility to use both word identification skills flexibly as the situation requires. That is, poor readers are characterized as inflexible in their over reliance on one or the other single word identification skill. This leads to the two major predictions of the current thesis. First, it is predicted that poor readers can be divided into two groups, whole-word poor readers and recoding poor readers respectively. This is demonstrated in Study 1 of the current thesis through an examination of the pattern of correct and incorrect responses to word lists in both the reading and spelling of poor readers in grade 3. The outcome of Study 1 is two groups of poor readers who rely on one or the other word identification skill. Experiments 1 and 2 demonstrate that the recoding poor readers are more sensitive to sound-based properties of words than whole-word poor readers although Experiment 3 shows that this is true only for the purpose of lexical access, not for the purpose of promoting efficient use of S.T.M. Experiments 4 and 5 demonstrate that whole-word poor readers rely more heavily on the overall visual appearance of words than recoding poor readers. Second, it is predicted that poor readers should show no strong deficit in the use of higher-order information to aid in word identification. Studies 2 and 3 demonstrate that both groups of poor readers can use context to aid in word identification. Study 2 demonstrates this for a simple category context and Study 3 demonstrates this for text. The results of Study 2 and 3 also show that contextual information overshadows the information of single word identification skills. That is, clear differences in skill preferences between the two types of poor reader when random word lists are read disappear when contextually constrained situations are used. This leads to the argument that contextually constrained "natural" reading situations may not be optimal for the investigation of reading deficits.
Mitterer, John Otto, "THERE ARE TWO KINDS OF POOR READER" (1980). Open Access Dissertations and Theses. Paper 2939.