Does third grade discrepancy status predict the course of reading development?

Annals of Dyslexia, 2001 by Flowers, Lynn, Meyer, Marianne, Lovato, James, Wood, Frank, Felton, Rebecca

There is persisting debate concerning the use of an ability-achievement discrepancy formula to define and identify learning disabled-including reading disabled-students. This study employs mixed effects regression growth curve analysis to assess the developmental course of discrepant and nondiscrepant readers (within poor readers) who were identified in third grade and retested in fifth, eighth, and twelfth grades. The results showed that discrepancy status does not differentiate the developmental course of basic reading skills (word identification or decoding), reading comprehension, or underlying cognitive abilities (phonemic awareness and fluency) in poor readers. The ability-achievement discrepancy model is not supported. Educational and legislative reasons for the persisting difficulties of poor readers are explored and recommendations for changes in public policy are made.

Longitudinal outcome research has challenged the original assumption that reading difficulties were simply due to a maturational lag and that over time, these students catch up with their peers (Satz, Taylor, Friel, & Fletcher, 1978). Lyon (1996) cites findings from NICHD studies showing that poor readers, identified in elementary school, do not catch up; rather, their reading problems persist through eighth and ninth grades. That reading difficulties persist, and that there are concomitant underlying cognitive weaknesses, is taken as strong evidence for a deficit model; that is, specific congenital cognitive weaknesses are the basis for the deficit. Given a deficit model hypothesis, the next question is whether general ability as measured by IQ tests predicts a differential outcome for poor readers. This study addresses the question by examining the developmental course of several reading and reading-related skills in students originally identified as IQ discrepant or nondiscrepant poor readers in third grade and following them through twelfth grade.

In the 1970s, when concerns about reading disabilities increased, there was a prevailing view that children with average or above-average ability had the potential to learn and that their failure to learn was both unexpected and unexplained. In contrast, children with less ability, described variously as "slow learners" or as "garden variety" poor readers (Gough & Tunmer, 1986), could not be expected to learn as well because, in this view, their potential itself was compromised. Indeed, the 1977 U.S. Office of Education definition of learning disabilities that included "severe achievement-ability discrepancy" as a major component of learning disability criteria formalized that assumption.

Over the next three decades, research studies converged to reveal the underlying cognitive abilities needed to acquire reading skills. Well-replicated research has demonstrated that a core deficit for reading disabled individuals-both children and adults-is phonemic awareness, the ability to understand how sounds and sound patterns work in our language system (Felton, Naylor & Wood, 1990; Liberman, Shankweiler, & Liberman, 1989; Stanovich & Siegel, 1994). Although phonemic awareness and phonological decoding are necessary prerequisites to efficient reading, they are not considered to be sufficient (Adams, 1990; Juel, Griffith, & Gough, 1986). Indeed, subsequent research has highlighted the role of other factors such as orthographic processing (Badian, 1995), short-term auditory memory (Torgeson, 1994), and particularly fluency (Denckla & Rudel, 1976; Wolf, Bally, & Morris, 1986).

As these underlying cognitive abilities were identified, researchers considered whether IQ-achievement discrepant ("underachievers") and nondiscrepant ("garden variety") students exhibited the same deficits, a test of whether slow learners are different in meaningful ways from their counterparts of average ability. Rutter and Yule (1975) suggested that students with higher verbal ability have more severe nonword reading deficits. While Stanovich (1988) initially found evidence that students whose reading is poor in spite of average verbal intelligence show more severe and more focal deficits in nonword reading than children whose verbal skills are lower, in a later article (Stanovich & Siegel, 1994) he found no such difference when using a new statistical model of regression-based logic. Other researchers (Badian, 1996, 2000; Biddle, 1996; Wolf, 1997) also have reported consistently a higher proportion of multiple deficits in underlying abilities or cognitive subprocesses (phonemic awareness, naming speed, and orthographic skills) among discrepant poor readers. However, the weight of the evidence is to the contrary, especially when more comprehensive longitudinal assessments have been used. Thus, using regression-based individual growth curve analysis, Stanovich and Siegel (1994), Fletcher, et al., (1994), Francis, Stuebing, Shaywitz, Shaywitz, and Fletcher (1996), and Foorman, Francis, Fletcher, and Lynn (1996) found no differences between IQ discrepant and nondiscrepant readers on a variety of underlying abilities such as word recognition, short-term verbal and nonverbal memory, vocabulary knowledge, and orthographic memory. Importantly, there has been general agreement across investigators that phonemic awareness is a core deficit for reading disabled individuals, whether discrepant or nondiscrepant.