This study examined factors related to survey response rate, particularly for teachers who participated in a nationwide survey. Using a newly developed statistical technique, the classification tree algorithm (CART), this study classified the lowest response rate and highest response rate groups based on their school demographic characteristics. It also examined the differences in teacher satisfaction, decision making autonomy, and classroom climate between the two groups of responders. The teachers participated in the High Performance Learning Communities (HiPlaces) Assessment in 1996-1997, an assessment that was designed to examine the degree to which a broad range of recommendations for effective school reform were implemented in a school and the impact of these reform efforts. In 1996-1997, 207 schools in 15 states participated in HiPlaces, and response data were available for 142 schools. The percentage of students eligible for free lunch appeared to be the most important factor among demographic variables in classifying schools with the lowest and highest response rates. In general, schools with the higher percentage of free lunch students have lower staff response rate, but the relationship is not linear. The CART approach allowed the examination of the nonlinear relationship. School size and the year the school joined the initiative were also important factors explaining the highest and lowest response rate groups. Systematic differences regarding the quality of the school as a work place and as an educational setting were found between the highest and lowest response rate groups. Teachers in the highest response rate group were more satisfied with their work, had less role conflict, and reported fewer barriers to implementation of the reform practices. The cumulative pattern of these findings suggests that response rates from staff tend to be lower in the schools that are of particular concern to educational reform efforts. (Contains 2 tables, 1 chart, and 10 references.) (SLD)