Evidence on the predictive validity of the Vane Kindergarten Test (VKT) as a part of the comprehensive screening and assessment battery administered by the Horseheads Central School District in New York is presented. Variable numbers of students at each of six elementary schools were administered the VKT. At the end of the kindergarten year, all students were administered the Stanford Early School Achievement Test, Level I as a part of the district's regular testing program. Individual regression equations and Pearson product-moment correlations were calculated for each member school and for the district. The full scale Vane IQ was correlated with total test raw score on the Stanford. The correlations are presented by school building. The three schools with the highest standard deviations in achievement scores were those experiencing the highest predictive validity. Credence is therefore added to the proposition that the low validities could be attributable to truncated or range anomalies and not to the test itself. At the present time, the relationship at the end of kindergarten is sufficiently great to warrant a combination of the Vane IQ with other variables, and does not suggest abandoning the test at Horseheads. (RC)