This research is part of a larger, IES-funded study titled: "Measuring the Efficacy and Student Achievement of Research-based Instructional Materials in High School Multidisciplinary Science" (Award # R305K060142). The larger study seeks to use a cluster-randomized trial design, with schools as the unit of assignment, to make causal inferences about the effect of treatment on both students and teachers. The research described in this report addresses the following research question associated with path "a" in Figure 1: (1) What is the mean difference in teacher outcome (i.e., instruction) across the treatment groups? (a) What is the effect size (practical significance)? (b) Is the difference statistically significant at the alpha = 0.05 level?; and (2) If practically or statistically significant differences in instruction exist across treatment groups, to what extent can the differences be attributed to the treatment (instructional materials and PD)? The research takes place in both suburban and rural high schools in the state of Washington. In particular, the suburban schools are clustered near Seattle/Tacoma and the rural schools are clustered near Yakima. The data from this analysis suggest that the PD treatment was more effective in fostering reform-oriented science instruction, on average, than was the extant PD experienced by the business-as-usual comparison group. This difference was both statistically and practically significant. Applying this result to the authors' hypothesis of mediation, they now have confidence that one of the causal paths (path a) that are necessary to argue mediation is trustworthy. Further study of path b is necessary to understand whether instruction is serving as a mediator of the treatment effect. That said, there is evidence in the literature suggesting that the possibility of a significant b path is quite real. For example, Hedges and Hedberg (2007) found that in school-level interventions, a considerable amount of the variance in outcomes was attributable to teacher and /or classroom effects. Threats to internal validity that are noteworthy include limitations in the authors' confidence that the post-intervention differences in RTOP scores were not pre-existing (i.e., not attributable to the treatment). Unfortunately, they did not have a baseline RTOP measure that could have served as a covariate in the main effect analysis of treatment. Use of such a covariate would have likely provided a more precise estimate of the treatment effect. Further, because the comparison group received business-as-usual PD, this experience was highly variable across teachers. The research team has only cursory knowledge of the nature and duration of extant PD experienced by the comparison group. As such, there is limited clarity in the PD experiences to which the treatment is being compared. In the context of an efficacy trial, external validity (i.e., generalizability) of findings is not paramount. However, it should be noted again that the authors' sampling approach was not random. Therefore, they are cautious not to suggest that their treatment effect estimates would generalize far beyond their sample of rural and suburban schools in Washington state. (Contains 1 figure and 6 tables.)