Purpose-In order to support evidence-based practice, this study served to evaluate the diagnostic accuracy, convergent validity, and divergent validity of the Structured Photographic Expressive Language Test-Preschool: Second Edition (SPELT-P2; J. Dawson, J. A. Eyer, J. Fonkalsrud, 2005) in order to determine whether it can be used as a valid measure for identifying language impairment in preschoolers.Method-The SPELT-P2 was administered to 54 children with typically developing language and 42 children with specific language impairment.Results-A discriminant analysis revealed good sensitivity (90.6%), good specificity (100%), and good positive and negative likelihood ratios, with a standard score cutoff point of 87 used to determine group membership. Analyses of convergent and divergent validity also supported use of the SPELT-P2 for identifying language impairment in preschoolers.Implications-The empirical evidence supports use of the SPELT-P2 as a valid measure for correctly identifying the presence or absence of language impairment in 4-and 5-year-old preschool children. Keywords diagnostics; evidence-based practice; language; specific language impairment Norm-referenced tests are frequently used when diagnosing young children with language impairment. Because of the social, academic, monetary, and ethical consequences of such a diagnosis, it is critical to ensure that the tests that are designed to determine the presence or absence of a language impairment are indeed valid for that purpose (Anastasi, 1988;Messick, 1989;Plante & Vance, 1994;Spaulding, Plante, & Farinella, 2006;Zhang & Tomblin, 2000). Since 1997, the Individuals With Disabilities Education Act (IDEA, 1997, P.L. 105-17) has echoed this necessity, requiring that assessments and evaluation measures be "used for purposes for which [they] are valid and reliable" (118 Stat. 2705). This concept was retained in subsequent versions of this law (e.g., Individuals with Disabilities Education Improvement Act of 2004; P.L. 108-446, Stat. 2705).As a preliminary step, many test makers consult a panel of experts to evaluate whether the content of the test appears representative of the phenomenon being assessed (content validity). NIH-PA Author ManuscriptNIH-PA Author Manuscript NIH-PA Author ManuscriptHowever, this process is insufficient to establish the extent to which a test reflects its underlying construct (construct validity) (Gray, Plante, Vance, & Henrichsen, 1999;Messick, 1989). Instead, Dollaghan (2004) advocated an evidence-based practice framework that relies on credible studies to assess diagnostic assessments and procedures. Thus, before using a test for diagnostic purposes, the test's construct validity should be established through empirical evaluation.The primary evidence needed to demonstrate that a test can be used to identify cases of language impairment is its diagnostic accuracy. An evaluation of diagnostic accuracy must prove that the test is able to adequately identify children with language impairment as having language im...

No abstract

Purpose This randomized controlled study investigated whether a supplemental treatment designed to teach triadic gaze (TG) as a signal of coordinated joint attention (CJA) would yield a significantly greater increase in TG in the experimental versus control group. Method Eighteen 10- to 24-month-old children with severe motor impairments were randomly assigned to an experimental (n=9) or control group (n=9). For approximately 29 sessions over 17 weeks, experimental participants received TG treatment twice weekly with a speech-language pathologist (SLP) in addition to standard practice. Controls received only standard practice from birth-to-three therapists. Coders masked to group assignment coded TG productions with an unfamiliar SLP at baseline, every three weeks during the experimental phase, and at the final measurement session. Results TG increased across groups from baseline to final measurement, with the experimental group showing slightly greater change. Performance trends were examined using experimental phase moving averages. Comparisons revealed significant differences between groups at two time points (at 12 weeks, r= .30, a medium effect and at the end of the phase r=.50, large effect). Conclusion Results suggest the promise of a short-term, focused treatment to teach TG as a behavioral manifestation of CJA to children with severe physical disabilities.