Estimating Dynamic Treatment Effects from Project STAR (Working Paper 36)

This paper considers the analysis of data from randomized trials which offer a sequence of interventions and suffer from a variety of problems in implementation. In experiments that provide treatment in multiple periods (T>1), subjects have up to 2^{T}-1 counterfactual outcomes to be estimated to determine the full sequence of causal effects from the study. Traditional program evaluation and non-experimental estimators are unable to recover parameters of interest to policy makers in this setting, particularly if there is non-ignorable attrition. We examine these issues in the context of Tennessee's highly influential randomized class size study, Project STAR. We demonstrate how a researcher can estimate the full sequence of dynamic treatment effects using a sequential difference in difference strategy that accounts for attrition due to observables using inverse probability weighting M-estimators. These estimates allow us to recover the structural parameters of the small class effects in the underlying education production function and construct dynamic average treatment effects. We present a complete and different picture of the effectiveness of reduced class size and find that accounting for both attrition due to observables and selection due to unobservable is crucial and necessary with data from Project STAR

UNDERSTANDING ESSAY RATING AS A SOCIALLY MEDIATED ACTIVITY: THE CASE OF A HIGH-STAKES ENGLISH TEST IN CHINA

Most essay rating research in language assessment has examined human raters’ essay rating as a cognitive process, thus overlooking or oversimplifying the interaction between raters and sociocultural contexts. Given that raters are social beings, their practices have social meanings and consequences. Hence it is important to situate essay rating within its sociocultural context for a more meaningful understanding. Drawing on Engeström’s (1987, 2001) cultural-historical activity theory (CHAT) framework with a sociocultural perspective, this study reconceptualized essay rating as a socially mediated activity with both cognitive (individual raters’ goal-directed decision-making actions) and social layers (raters’ collective object-oriented essay rating activity at related settings). In particular, this study explored raters’ essay rating at one provincial rating centre in China within the context of a high-stakes university entrance examination, the National Matriculation English Test (NMET). This study adopted a multiple-method multiple-perspective qualitative case study design. Think-aloud protocols, stimulated recalls, interviews, and documents served as the data sources. This investigation involved 25 participants at two settings (rating centre and high schools), including rating centre directors, team leaders, NMET essay raters who were high school teachers, and school principals and teaching colleagues of these essay raters. Data were analyzed using Strauss and Corbin’s (1990) open and axial coding techniques, and CHAT for data integration. The findings revealed the interaction between raters and the NMET sociocultural context. Such interaction can be understood through a surface structure (cognitive layer) and a deep structure (social layer) concerning how raters assessed NMET essays, where the surface structure reflected the “what” and the deep structure explained the “how” and “why” in raters’ decision-making. This study highlighted the roles of goals and rules in rater decision-making, rating tensions and raters’ solutions, and the relationship between essay rating and teaching. This study highlights the value of a sociocultural view to essay rating research, demonstrates CHAT as a sociocultural approach to investigate essay rating, and proposes a direction for future washback research on the effect of essay rating. This study also provides support for NMET rating practices that can potentially bring positive washback to English teaching in Chinese high schools.