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.

Statistical methods for characterizing similarities and differences between semantic structures

This paper describes a variety of statistical methods for obtaining precise quantitative estimates of the similarities and differences in the structures of semantic domains in different languages. The methods include comparing mean correlations within and between groups, principal components analysis of interspeaker correlations, and analysis of variance of speaker by question data. Methods for graphical displays of the results are also presented. The methods give convergent results that are mutually supportive and equivalent under suitable interpretation. The methods are illustrated on the semantic domain of emotion terms in a comparison of the semantic structures of native English and native Japanese speaking subjects. We suggest that, in comparative studies concerning the extent to which semantic structures are universally shared or culture-specific, both similarities and differences should be measured and compared rather than placing total emphasis on one or the other polar position.

Social stress and the reactivation of latent herpes simplex virus type 1

Psychological stress is thought to contribute to reactivation of latent herpes simplex virus (HSV). Although several animal models have been developed in an effort to reproduce different pathogenic aspects of HSV keratitis or labialis, until now, no good animal model existed in which application of a psychological laboratory stressor results in reliable reactivation of the virus. Reported herein, disruption of the social hierarchy within colonies of mice increased aggression among cohorts, activated the hypothalamic-pituitary-adrenal axis, and caused reactivation of latent HSV type 1 in greater than 40% of latently infected animals. However, activation of the hypothalamic-pituitary-adrenal axis using restraint stress did not activate the latent virus. Thus, the use of social stress in mice provides a good model in which to investigate the neuroendocrine mechanisms that underlie behaviorally mediated reactivation of latent herpesviruses.

Detection of cortical activation during averaged single trials of a cognitive task using functional magnetic resonance imaging

Functional neuroimaging studies in human subjects using positron emission tomography or functional magnetic resonance imaging (fMRI) are typically conducted by collecting data over extended time periods that contain many similar trials of a task. Here methods for acquiring fMRI data from single trials of a cognitive task are reported. In experiment one, whole brain fMRI was used to reliably detect single-trial responses in a prefrontal region within single subjects. In experiment two, higher temporal sampling of a more limited spatial field was used to measure temporal offsets between regions. Activation maps produced solely from the single-trial data were comparable to those produced from blocked runs. These findings suggest that single-trial paradigms will be able to exploit the high temporal resolution of fMRI. Such paradigms will provide experimental flexibility and time-resolved data for individual brain regions on a trial-by-trial basis.

Soft-tissue characters in higher primate phylogenetics

Recent research has cast doubt on the reliability of bones and teeth for reconstructing phylogenetic relationships among higher primate species and genera. Herein, we investigate whether this problem is confined to hard tissues by examining the utility of higher primate soft-tissue characters for reconstructing phylogenetic relationships at low taxonomic levels. We use cladistic methods to analyze 197 soft-tissue characters for the extant hominoids and then compare the resulting phylogenetic hypotheses with the group's consensus molecular phylogeny, which is widely considered to be accurate. We show that the soft-tissue characters yield robust phylogenetic hypotheses that are compatible with the molecular phylogeny. Given the strength of the evidence for molecular phylogeny, these results indicate that, unlike craniodental hard-tissue characters, soft tissues are reliable for reconstructing phylogenetic relationships among higher primate species and genera. Thus, in higher primates at least, some types of morphological data are more useful than others for phylogeny reconstruction.

On a psychophysical transformed-rule up and down method converging on a 75% level of correct responses

Tranformed-rule up and down psychophysical methods have gained great popularity, mainly because they combine criterion-free responses with an adaptive procedure allowing rapid determination of an average stimulus threshold at various criterion levels of correct responses. The statistical theory underlying the methods now in routine use is based on sets of consecutive responses with assumed constant probabilities of occurrence. The response rules requiring consecutive responses prevent the possibility of using the most desirable response criterion, that of 75% correct responses. The earliest transformed-rule up and down method, whose rules included nonconsecutive responses, did not contain this limitation but failed to become generally accepted, lacking a published theoretical foundation. Such a foundation is provided in this article and is validated empirically with the help of experiments on human subjects and a computer simulation. In addition to allowing the criterion of 75% correct responses, the method is more efficient than the methods excluding nonconsecutive responses in their rules.