The discriminative efficiency of the Bayley Scales of infant development

This sudy was designed to determine the discriminative efficiency of the Bayley Mental and Motor Scales for classifying infants as neurologically suspicious and neurologically normal. The two groups employed in the study were formed on the basis of the results from comprehensive neurological examinations performed at ages 1 and 7. Three discriminant function analyses were conducted, one for each scale separately and a third incorporating both scales using a stepwise procedure. Efficiency was measured in terms of number of correct and incorrect classifications, false positive and false negative errors, and validity coefficients. Comparisons among the analyses were examined by means of relative discriminative power and incremental validity. The findings suggested that the Motor Scale provides the most accurate identification of infants with suspected neurological impairments. A statistically significant likelihood discriminant function derived from both scales yielded some additional discriminative power. The law of parsimony, however, dictated the use of the single Motor Scale.The author is grateful to Drs. Janet Hardy and Doris Welcher for providing the data upon which this research was based and to Dr. Patricia Gold for collecting the data. The data were gathered for the Johns Hopkins Collaborative Perinatal Project and supported by the National Institutes of Health (NINDS), the U.S. Children's Bureau through the Maryland State Department of Health, the American Academy of Ophthalmology and Otolaryngology, the Thomas Wilson Sanitarium, the Joseph P. Kennedy, Jr., Fluid Research Fund, The Johns Hopkins University, The Johns Hopkins Hospital, and the facilities of The Johns Hopkins Medical Computing Center. Anonymous reviewers are also acknowledged for their helpful suggestions in improving the overall clarity of the paper.