Two problems were considered in this study: (1) to determine if there were combinations of variables in the population studied that were predictive of science-teaching practices and/or teacher satisfaction with teaching science; and (2) to determine if models could be developed to explain possible directional effects of selected teacher variables on science-teaching practices and/or teacher satisfaction with teaching science. The population for the study consisted of all the public elementary schools in the New England Region, the Mideast Region, and the Southwest Region of the United States. The sample consisted of 2,948 randomly chosen public elementary schools. Data were analyzed by descriptive analysis consisting of frequency counts, percentages, means and standard deviations. Correlational analysis consisted of Pearson-Product Moment Correlation Coefficients between teacher variables. Predictive analysis using step-wise regression procedure and factor analysis were used. Teacher variables most frequently correlated with science teaching practices were: (1) science in-service participation, (2) availability of supplies and equipment. Best predictors of science-teaching practices were: (1) grade level, (2) limiting teacher effectiveness factors, and (3) availability of supplies. (Author/EB)