This study examines variables that affect staff development in rural schools. It examines systematically delivered rural staff development intervention in 28 rural schools in five states representative of the Southwest--Arkansas, Louisiana, Oklahoma, New Mexico, and Texas. The study, involving 700 teachers, tests 18 hypotheses based largely on research, experience, and literature from non-rural settings. It examines 24 variables related to staff development intervention, in order to construct a predictive model of context variables supporting the use of effective strategies for staff training. Initial site variables tested were: average daily attendance, ethnicity, socioeconomic status (SES), number of teachers and administrators, selection of site, proximity to the nearest institution of higher learning (IHE), and a perceptual measure of site's instructional focus. Training variables included: administrative participation, incentives offered, time of day held, release time allowed, number of participants, source of trainer, and training observation. Teacher rate of implementation was measured with the following variables: method of selection for training, gender, grade level, and type of course taught. The study concludes that a number of variables have an effect on the rate that teachers implement new or improved strategies and behaviors following training. Eight predictors were found to account for 56.9% of dependent variable, including: (1) training time of day; (2) number of students at school eligible for free or reduced lunch (SES); (3) administrative participation; (4) nearness to an IHE; (5) incentive provided; (6) high school-level teaching assignment; (7) number of administrators in district; and (8) trainer from an IHE. The study recommends areas of further research. Approximately 100 references are included. (TES)