The assistance of John Holahan, who bore responsibility for maintaining and updating the database, is gratefully acknowledged. Thanks are also due to Lois Dreyer, who supervised the testing of the children. In addition, we acknowledge many others at the Yale University Center for the Study of Learning and Attention Disorders and at Haskins Laboratories, whose help was essential to the completion of the project. Finally, we thank Linda Kimbrough for her work on preparation of this article.

A comprehensive cognitive appraisal of elementary school children with learning disabilities showed that within the language sphere, deficits associated with reading disability are selective Phonological deficits consistently accompany reading problems whether they occur in relatively pure form or in the presence of coexisting attention deficit or arithmetic disability Although reading-disabled children were also deficient in production of morphologically related forms, this difficulty stemmed in large part from the same weakness in the phonological component that underlies reading disability In contrast, tests of syntactic knowledge did not distinguish reading-disabled children from those with other cognitive disabilities, nor from normal children after covarying for intelligence

No abstract

Requests for reprints should be sent to Donald Shankweiler, Haskins Laboratories, 270 Crown Street, New Haven, CT 065 1 1. Downloaded by [McMaster University] at 10:09 03 November 2014 70 SHANKWEILER ET AL.Comparisons of reading measures from a sample of 361 children aged 7.5 to 9.5, including many with reading difficulties, showed high correlations between word reading and nonword reading, and between each of these abilities and reading comprehension. These results, together with other findings from thesechildren, showed that skill in word identification was almost inseparable from the phonologically analytic decoding process that is tapped by nonword reading, and, correspondingly, differences in reading comprehension were closely associated with differences in decoding skill. The findings support the conclusion that bottom-up skills largely drive the reading process in this age group. Only a small number of children departed from the norm in showing better reading comprehension than would be expected from their decoding skills, and those with the opposite discrepancy accounted for even fewer.

Mean length of utterance (MLU) in morphemes was examined as a predictor of the grammatical complexity of natural language corpora of normal preschoolers and of children and adolescents with delayed language, Fragile X syndrome, Down syndrome, and autism. The Index of Productive Syntax (IPSyn) served as the measure of syntactic and morphological proficiency. For all groups, a strong curvilinear association between measures was found across the MLU range from 1.0 to about 4.5. Correlations were weaker when MLU exceeded 3.0 than during earlier stages of language development, however, confirming previous suggestions that MLU becomes less closely associated with grammatical development as linguistic proficiency increases. For the language-disordered groups, moreover, the curves relating the two measures differed from the curves for the normal preschoolers because MLU frequently overestimated actual IPSyn scores. The results are discussed with respect to the use of MLU in conjunction with other measures of syntactic complexity in the study of atypical language development.