Preschool screening referral rates for developmental problems were examined for 219 Minnesota school districts across 2 school years. The state as a whole was consistent from year 1 to year 2 in terms of the percentage of children referred for screening. However, there was wide variation among individual districts. Three groups of districts were identified on the basis of the change in rates: (1) those whose rates increased considerably, (2) those whose rates decreased considerably, and (3) those whose rates stayed the same from year 1 to year 2. An attempt was made to identify variables that differentiated the groups. No differences were found among the three groups and the state as a whole, when compared on the basis of three district variables and four socioeconomic status variables. In addition, no differences were found in terms of the screening practices employed by these groups. The need for further examination of factors that contribute to referral rate changes is discussed. Implications of the wide referral rate variation for individual children and for screening efficiency in general are also discussed. (Author/CL)