Children whose language development is not proceeding at the expected rate, but who have no cognitive deficit or other developmental disorder, are considered to have a specific language impairment (SLI). The word specific refers to the supposedly circumscribed nature of the deficits found in SLI: they do not have sensory, cognitive, neurological, or socio-emotional impairments that might account for their language problems. Leonard (1998) describes the history of research and clinical interest in this group of children. Although terms and definitions have changed over the years, one constant has been that the language performance of children with SLI is not commensurate with their nonverbal intellectual ability.The idea of ruling out cognitive impairments in SLI has become instantiated in an informal but widely accepted standard of requiring children to earn a nonverbal IQ score no more than 1 SD below the mean (i.e., ≥ 85). Plante (1998) traced this custom back to the work of Stark and Tallal (1981), but noted that the context of the criterion in that study has been lost. Plante also noted that, although the nonverbal IQ criterion is intended to provide comparability of samples across studies, a score of 85 on one IQ test is not necessarily equivalent to an 85 on another test. In the current paper, we provide further evidence that nonverbal IQ scores are dependent on the test being used, and suggest that inter-test differences are not the same for all children.Tests that are intended to measure nonverbal IQ use a variety of tasks and theoretical constructs, and also vary in their psychometric properties (DeThorne & Schaefer, 2004). Thus, different nonverbal IQ tests not only attempt to measure different abilities, but vary in the reliability and validity with which they do so. It is to be expected that a single individual will not receive the same score on two nonverbal IQ tests if only because of measurement error, but in addition, the underlying abilities being evaluated are likely to be different for any two tests.Despite these issues, one purpose of applying nonverbal IQ criteria for SLI is to ensure that the scores of children with SLI and their typically developing peers are similar. Swisher and colleagues, however, have shown that even when similar scores are obtained on the same test, one cannot be certain that nonverbal abilities are truly equivalent. Swisher and Plante (1993) administered two nonverbal intelligence tests to children with and without SLI aged 4 to 5 years, and examined the covariance relationships among subtests, or groups of similar items, for each test. They found that these relationships were different for the SLI and normal language groups. In another study, Swisher, Plante, and Lowell (1994) administered three nonverbal IQ tests to children between the ages of 8 and 10 years. The results showed that scores on the tests Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers