Academic journal article School Psychology Review

Individual Part Score Profiles of Children with Intellectual Disability: A Descriptive Analysis across Three Intelligence Tests

Academic journal article School Psychology Review

Individual Part Score Profiles of Children with Intellectual Disability: A Descriptive Analysis across Three Intelligence Tests

Article excerpt

School psychologists often rely on intelligence tests to provide necessary information in order to make decisions regarding eligibility for special education services and placement of students in appropriate educational settings (Kamphaus, 2009). In addition to providing an IQ, which is almost always the most reliable score yielded by intelligence tests and the best predictor of educational outcomes, intelligence tests also yield part scores that stem from subsets of the same subtests that contribute to the IQ. These part scores may be called indexes, clusters, or factor scores. In general, part scores labels (e.g., Verbal Comprehension [Wechsler, 2003] and Fluid Reasoning [Kaufman & Kaufman, 2004]) more transparently reflect the item content and the response processes involved in completing their constituent subtests than do IQs that measure global ability. Thus, interpretation of part scores seems to provide a more logical link between test scores and instructional interventions (Jensen, 1984). Perhaps for these reasons, school psychologists have long interpreted part score profiles by examining both cognitive strengths and weaknesses, and IQs (Kamphaus, Winsor, Rowe, & Kim, 2012).

During the past several decades, there have been two notable trends in part score interpretation that are based on both pragmatic and theoretical opposition to interpreting only IQs. First, test users consider IQs in interpretive models that focus on examination of score hierarchies through successive levels of interpretation. For example, Kaufman (1994), Groth-Marnat (1997), and Sattler (2008) recommended that the "validity" of the IQs of individuals be questioned when there is substantial variability among part scores that compose the IQ. Thus, these models guide practitioners to disregard the IQ in cases of significant differences in part scores and to focus subsequently on part score profiles.

Second, following the belief that interpreting the IQ is limiting because it fails to adequately represent the breadth of human cognitive abilities, some scholars have asserted that the purpose of intelligence testing is to interpret part scores and to largely ignore the IQ (e.g., Fiorello et al., 2007; Horn & Blankson, 2005; McGrew & Flanagan, 1998). Consistent with this these assertions, the number and variety of part scores yielded by intelligence tests have increased in recent years (Frazier & Youngstrom, 2007), and guidelines for assessing a broad range of specific cognitive abilities via parts scores have been offered (e.g., Flanagan, Ortiz, & Alfonso, 2012). In particular, part score interpretation based on the Cattell-Horn-Carroll (CHC) theory of cognitive abilities appears to have proliferated (Newton & McGrew, 2010).

Considerations for Part Score Interpretation

Despite the widespread practice part score interpretation, either in addition to IQs or in place of them, the scientific evidence base for doing so is ambiguous at best. Whereas some researchers have found evidence that part scores provide unique contributions in predicting the IQ for children with various disabilities (e.g., Fiorello et al., 2007; Fiorello, Hale, McGrath, Ryan, & Quinn, 2001), others have found, for example, that the IQ remains a strong predictor of achievement outcomes even though part score variability and interpretation of part scores provides no additional meaningful information beyond IQ (e.g., Daniel, 2007; Freberg, Vandiver, Watkins, & Canivez, 2008; Watkins, Glutting, & Lei, 2007). Furthermore, consistent patterns of part score profiles for high-incidence disability groups have not emerged (Kaufman, 2000).

There are also psychometric limitations that must be considered when interpreting part scores. First, part scores inherently have lower reliabilities than associated IQs because part scores are based on far fewer constituent subtests than the overall IQ. Thus, part scores vary more from measurement error than do the more reliable IQs; for all scores, measurement error is a persistent undermining influence on the validity of interpretations (Jensen, 1984). …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.