Concern with the problem of science curriculum implementation has been highlighted by surveys which have shown that the impact of the major NSF sponsored curriculum projects has not been as pervasive as might be expected. Although there has been some recent research and theory development on the problem of implementation, little work has been done on the determinants of implementation, especially in science. This article describes a replicated study of factors influencing implementation of a hypothetical elementary science curriculum. The instrument used allowed teachers to respond to a particular implementation scenario devised by combining a number of statements about properties of an innovation with school and teacher reactions to the innovation. Responses were given as estimates of the probability of implementation under the conditions described in the scenario. Results showed that the staff attitude factor was by far the most salient influence on implementation. An inservice factor which was highly significant in one of the applications of the instrument failed to replicate. This suggested the possibility that context differences between the two samples may have been important for this factor. The findings are interpreted in relation to Fullan's model of implementation and to other research which highlights the importance of individual interpretation of the properties of an innovation.