One hundred and twenty, full-term, normal, first-born infants were categorized at one month of age into a high-risk or low-risk group for possible development of emotional and developmental deviations. The predictions were based on measurements of the mother's perception of her infant as compared to the average. At age 4 1/2, the children were evaluated by two child psychiatrists who had no knowledge of the children's predictive risk rating. A statistically significant association was evident between prediction and outcome. This paper describes the methodology and discusses the implications of the findings.