Dennis L. Molfese Victoria J. Molfese Leslie A. Gill Sharon Benshoff Southern Illinois University at Carbondale
Since the 1970s, researchers and practitioners have been interested in the development of assessment tools for neonates that are predictive of cognitive status in later infancy and early childhood. The ideal measures would permit the assessment of abilities at birth, when virtually total populations of infants are readily accessible in hospitals, and would be easily administered, cost effective, and accurate in identifying those infants who are at risk for developmental delays. However, the ideal measures that satisfy all of these criteria have been difficult to identify. Most typical approaches to assessment have involved the use of a wide variety of newborn measures as predictors and a variety of performance measures as the criterion scores. The newborn and early infancy measures used as predictors have included measures of perinatal complications (e.g., the Obstetrical Complications Scale, Littman & Parmelee, 1978), neurological and behavioral assessments (e.g., the Brazelton Neonatal Assessment Scale, Brazelton, 1973; Prechtl Neurological Examination, Prechtl, 1968), electrophysiological measures of brain functioning (e.g., brainstem-auditory evoked responses, evoked brain potentials), and measures reflecting attention and tactile abilities. Criterion measures have included scores on scales such as the Bayley Scales of Infant Development ( Bayley, 1969), the Denver Developmental Screening Test ( Dunn, 1965), the Stanford-Binet Intelligence Scale ( Thorndike, Hagen, & Sattler, 1986), and the McCarthy Scales of Children's Abilities ( McCarthy, 1972).
In the majority of studies that have been published, the amount of variance accounted for by the predictor variables, alone or in combination, is low. Since 1990, there has been an increase in the variance accounted for