Title

Phonological complexity, decoding, and text comprehension

Date of Completion

January 2003

Keywords

Psychology, Cognitive

Degree

Ph.D.

Abstract

Relationships between decoding and other reading abilities were investigated in experienced readers. Based on the distribution of decoding scores in a norming study with 5444 university students, 63 participants were selected, 30 high scorers and 33 low scorers. They were compared on measures of reading and related abilities: phonological awareness, nonword naming, spelling, working memory, print exposure, vocabulary, reading comprehension, reading rate, and oral reading of pseudotext. Significant differences were observed in phonological awareness, spelling, and accuracy but not speed of nonword and pseudotext reading, as well as print exposure and reading rate. ^ Two explanations of the relationship between decoding and comprehension differences were tested, each assuming a limited, common pool of cognitive resources that can be devoted to the component tasks of reading, ranging from word recognition through text comprehension. The Threshold Hypothesis emphasizes the importance of text content and structure, and predicts that when text difficulty exceeds some threshold, all processes that rely on working memory will be taxed to the point that efficiency of decoding limits comprehension outcomes. The Local Effects Hypothesis supposes that the basis of the difficulty lies in particular words in the text, predicting that text containing difficult-to-decode words will be comprehended less well than easily decoded text because more shared resources are diverted to the word-recognition process. Participants read two texts of equal difficulty, differing only in decoding demands. Consistent with the Threshold Hypothesis, no effect of decoding difficulty or participants' decoding skill was observed. ^ In the combined sample of participants, response time measures of basic skills (nonword reading and phonological awareness) accounted for 15% of the variance in reading rate. Pseudotext oral reading time explained an additional 13%, attributed chiefly to fluency in visual parsing of text. The time measures involving basic skills explained 12% of the variance in reading comprehension, and after controlling for general verbal ability, nonword reading time still accounted for significant variance in comprehension. The implications of phonological awareness and decoding differences for reading efficiency and comprehension are discussed in terms of phonological representations of lexical items that may vary in quality even among practiced readers. ^