A model for the development of scientific reasoning in adolescents was formulated largely upon the basis of Piagetian theory. Included as potential determinants of scientific reasoning were: experience; age; locus of control; field dependence-independence (FID); rigidity/flexibility; intelligence quotient (IQ); and sex. Causal relationships between these variables were hypothesized a priori with strong theoretical, heuristic, and empirical support. Data (obtained from 106 middle school students and 92 high school students) were used to test the hypothesized model by path analysis. Multiple regressions were performed according to path analytic methods to acquire standardized beta weights for each of the hypothesized paths. These beta values were used as path coefficients for each of the posited relationships. Significant coefficients were obtained for these variables and scientific reasoning: age; IQ; FID; and experience. Age and IQ were stronger determinants of scientific reasoning that were FID and experience. An indirect effect of locus of control on scientific reasoning through the FID variable was also supported. None of the paths involving sex or rigidity/flexibility was significant. The revised model (which excluded sex and rigidity/flexibility) included significant paths which explained 61 percent of the variance in scientific reasoning. Implications for science instruction are addressed. (JN)