Intrapolation and extrapolation of age-equivalent scores for the Bayley II: a comparison of two methods of estimation

The Bayley Scales of Infant Development II (1) is a well established standardized test for assessing the mental ability of infants and young children. It provides an age-adjusted standard score as a summary measure, but for very low (or very high) functioning children the raw scores on this test may not allow the calculation of a standard score. The manual provides for the transformation of raw scores into age-equivalents but this translation is not unique and results in a step function. The availability of a unique and continuous transformation of raw scores to age-equivalents is critical when evaluating longitudinal mental development, particularly in the environment of controlled clinical trials or natural history studies. We compared two methods for deriving unique age equivalents: a local regression method, with estimates restricted to age-equivalents within the age range of the test, and a linear method, which also allows extrapolation outside the age range of the test. The linear method was found to be most useful and the values, which are provided in table format, can be used for assigning age equivalent scores to individual children. They are also useful in clinical trials which use the Bayley to assess the safety and efficacy of treatments with potential cognitive effects, when conducted in populations where the children are likely to record scaled scores below the limits of the test.