This paper analyzes 12 recent within-study comparisons contrasting causal estimates from a randomized experiment with those from an observational study sharing the same treatment group. The aim is to test whether different causal estimates result when a counterfactual group is formed, either with or without random assignment, and when statistical adjustments for selection are made in the group from which random assignment is absent. We identify three studies comparing experiments and regression-discontinuity (RD) studies. They produce quite comparable causal estimates at points around the RD cutoff. We identify three other studies where the quasi-experiment involves careful intact group matching on the pretest. Despite the logical possibility of hidden bias in this instance, all three cases also reproduce their experimental estimates, especially if the match is geographically local. We then identify two studies where the treatment and nonrandomized comparison groups manifestly differ at pretest but where the selection process into treatment is completely or very plausibly known. Here too, experimental results are recreated. Two of the remaining studies result in correspondent experimental and nonexperimental results under some circumstances but not others, while two others produce different experimental and nonexperimental estimates, though in each case the observational study was poorly designed and analyzed. Such evidence is more promising than what was achieved in past within-study comparisons, most involving job training. Reasons for this difference are discussed. © 2008 by the Association for Public Policy Analysis and Management.

This paper analyzes 12 recent within-study comparisons contrasting causal estimates from a randomized experiment with those from an observational study sharing the same treatment group. The aim is to test whether different causal estimates result when a counterfactual group is formed, either with or without random assignment, and when statistical adjustments for selection are made in the group from which random assignment is absent. We identify three studies comparing experiments and regression-discontinuity (RD) studies. They produce quite comparable causal estimates at points around the RD cutoff. We identify three other studies where the quasi-experiment involves careful intact group matching on the pretest. Despite the logical possibility of hidden bias in this instance, all three cases also reproduce their experimental estimates, especially if the match is geographically local. We then identify two studies where the treatment and nonrandomized comparison groups manifestly differ at pretest but where the selection process into treatment is completely or very plausibly known. Here too, experimental results are recreated. Two of the remaining studies result in correspondent experimental and nonexperimental results under some circumstances but not others, while two others produce different experimental and nonexperimental estimates, though in each case the observational study was poorly designed and analyzed. Such evidence is more promising than what was achieved in past within-study comparisons, most involving job training. Reasons for this difference are discussed. © 2008 by the Association for Public Policy Analysis and Management. INTRODUCTIONComprehensive program evaluation depends on validly determining a program's causal impacts. Debate has been vigorous about the role experiments and observational studies should play in identifying such impacts. The main reason for preferring experiments is that, when perfectly implemented, they create intervention and control groups that do not initially differ in expectation and so do not differ on any measured or unmeasured variables. However, the regression-discontinuity design (RD) and instrumental variables (IV) also provide unbiased causal inference in theory. So additional technical justification for preferring experiments is required. It comes from experimental estimates being more precise than RD and IV estimates (Goldberger, 1972) and also from the experiment's assumptions being more transparent in research practice. IV's main assumption is that the instrument is only correlated with outcome through treatment. This assumption is well warranted when the treatment is allocated via random assignment (Angrist, Imbens, & Rubin, 1996) or RD (Hahn, Todd, & van der Klauuw, 2001), but otherwise it is usually unclear whether the assumption is met in practice. RD's main assumption is that the functional form relating the assignment variable to outcome is fully known. This can usually be checked quite well in the data, but perhaps not quite as well as with the ...

Since 1980, the number of state pre-kindergarten (pre-K)

This article evaluates whether providing coaching between practice sessions in teacher education courses leads to more rapid development of skills and changes in teachers’ beliefs about student behavior, using mixed-reality simulations as a practice space and standardized assessment platform. We randomly assigned 105 prospective teachers to different coaching conditions between simulation sessions integrated into a teacher preparation program. Coached candidates had significant and large improvements on skills relative to those who only reflected on their teaching. We also observe significant coaching effects on candidates’ perceptions of student behavior and ideas about next steps for addressing perceived behavioral issues. Findings suggest that skills with which novices struggle can improve with coaching and do not have to be learned “on the job.”

Sertoli--Sertoli and Sertoli--germ-cell configurational relationships were studied using morphometric techniques and direct measurements as obtained from micrographs used to reconstruct a model of a rat stage V Sertoli cell. Regional areas of the Sertoli cell surface, which faced germ cells, other Sertoli cells, or noncellular structures, were expressed as relative surface area percentages; and the absolute surface areas for these regional areas were calculated. The surface areas of the reconstructed cell, in its unmagnified state, was found to be 12,163 micron2. Cell processes were enumerated and studied using morphometric techniques. The surface area of the reconstructed Sertoli cell facing germ cells and Sertoli cells was also determined. Five Sertoli cells showed extensive contact with the reconstructed cell at the level of the Sertoli--Sertoli junctional contact region. This contact region averaged 3.51 micron in width. The relative and absolute surface area of subsurface ectoplasmic specialization of the Sertoli cell that faced germ cells and other Sertoli cells was calculated, and the extent of penetration of step 17 spermatids into the Sertoli crypts was determined. Surface relationships of the reconstructed cell to cellular and noncellular elements were depicted on outline drawings of the Sertoli cell.