Sources of selection bias in evaluating social programs: An interpretation of conventional measures and evidence on the effectiveness of matching as a program evaluation method

This paper decomposes the conventional measure of selection bias in observational studies into three components. The first two components are due to differences in the distributions of characteristics between participant and nonparticipant (comparison) group members: the first arises from differences in the supports, and the second from differences in densities over the region of common support. The third component arises from selection bias precisely defined. Using data from a recent social experiment, we find that the component due to selection bias, precisely defined, is smaller than the first two components. However, selection bias still represents a substantial fraction of the experimental impact estimate. The empirical performance of matching methods of program evaluation is also examined. We find that matching based on the propensity score eliminates some but not all of the measured selection bias, with the remaining bias still a substantial fraction of the estimated impact. We find that the support of the distribution of propensity scores for the comparison group is typically only a small portion of the support for the participant group. For values outside the common support, it is impossible to reliably estimate the effect of program participation using matching methods. If the impact of participation depends on the propensity score, as we find in our data, the failure of the common support condition severely limits matching compared with random assignment as an evaluation estimator.

Flower color variation: A model for the experimental study of evolution

We review the study of flower color polymorphisms in the morning glory as a model for the analysis of adaptation. The pathway involved in the determination of flower color phenotype is traced from the molecular and genetic levels to the phenotypic level. Many of the genes that determine the enzymatic components of flavonoid biosynthesis are redundant, but, despite this complexity, it is possible to associate discrete floral phenotypes with individual genes. An important finding is that almost all of the mutations that determine phenotypic differences are the result of transposon insertions. Thus, the flower color diversity seized on by early human domesticators of this plant is a consequence of the rich variety of mobile elements that reside in the morning glory genome. We then consider a long history of research aimed at uncovering the ecological fate of these various flower phenotypes in the southeastern U.S. A large body of work has shown that insect pollinators discriminate against white phenotypes when white flowers are rare in populations. Because the plant is self-compatible, pollinator bias causes an increase in self-fertilization in white maternal plants, which should lead to an increase in the frequency of white genes, according to modifier gene theory. Studies of geographical distributions indicate other, as yet undiscovered, disadvantages associated with the white phenotype. The ultimate goal of connecting ecology to molecular genetics through the medium of phenotype is yet to be attained, but this approach may represent a model for analyzing the translation between these two levels of biological organization.