Path analysis techniques were used to test a hypothesized structural model of direct and indirect causal effects of student variables on science process skills. The model was tested twice using data collected at the beginning and end of the school year from 67 9th- and 10th-grade biology students who lived in a rural Franco-American community in New England. Each student variable was found to have significant effects, accounting for approximately 80% of the variance in science process skills achievement. Academic ability, biology knowledge, and language preference had significant direct effects. There were significant mediated effects by cognitive development, parents' education, and attitude toward science in school. The variables of cognitive development and academic ability had the greatest total effects on science process skills. Implications for practitioners and researchers are discussed.