Problem. Current reform efforts in science education are constructivist in nature and call for major changes in the way science has been taught in the schools. Teacher efficacy, a measure of perceived instructional empowerment, is one variable which has been linked to teacher change and general classroom innovation. However, the specific relationship between efficacy and innovative science instruction had not been examined.

Method. This descriptive study employed a correlational design utilizing cross-sectional survey methodology. Data were collected via a three-part survey instrument. The purpose ofthis design was to gather descriptive data on science education in seventh- and eighth-grade classes in the United States and to correlate reported use of instructional practiceswith teacher efficacy scores and selected context variables.

Study participants totaled 285 from an original national sample of 543, for a net return rate of 52.5%. The data from these respondents were presented through descriptive statistics, Spearman rho correlation, and chi square.

Results. Data were presented concerning the reported use of specific instructional practices in seventh- and eighth-grade science education in the United States. Discussion and lecture were the two most commonly used instructional methods. Results show use of hands-on lab activities increased 4%, while use of lecture has decreased almost 6% since 1977.

Hypothesis testing resulted in the rejection of both of the study's null hypotheses. Significant relationships were found between 34 specific instructional practices and teacher efficacy. The number and size of correlation coefficients were greater between efficacy and constructivist practices than between efficacy and traditional, absorption-type practices, although all correlations were weak.

Fifty-three statistically significant correlations were found between use of specific instructional practices and years of teaching and between specific instructional practicesand perceived qualifications to teach science classes.

Conclusions. While these correlations were statistically significant, they were typically small. The descriptive data suggested the use of a variety of teaching practices by science teachers in the seventh- and eighth-grade classroom. The large number of small yet significant correlations supports this conceptualization. External validity of this sample was supported through a comparison of demographic features with Weiss's (1994) national probability sample.