The first 3 years of a 4-year project to plan and implement school change using a school improvement strategy in five low income, rural, resource-bound school sites are described in this report. One school from each of the following states--Texas, New Mexico, Oklahoma, Arkansas, and Louisiana--participated in the project. Extreme diversity characterized the schools in terms of population density, minority populations, climate and topography, natural resources, and economies. Quantitative data analysis examined reading, mathematics, and language arts test scores. Qualitative analysis developed contextual indicators, or criteria for improvement, unique to each school. Each school developed its own set of contextual indicators and demonstrated significant changes in school climate. Three schools reported significant improvements in test scores. Findings indicate that two sets of school improvement standards exist: universal ones for student achievement outcomes and those unique to a school's goals and objectives. Recommendations are made for the consideration of the whole school context in assessing school effectiveness and for measuring school improvement in a variety of ways. Four tables are included. (43 references) (LMI)