The use of regression equations to predict the expected achievement levels of five- and six-year-old children on the Wide Range Achievement Test are discussed. Problems are found in identifying underachievement for children in the lower primary grades. Implications for the early identification of specific learning disabilities are described. (Author)