A Bayesian Multilevel Modeling Approach to Time-Series Cross-Sectional Data

Joseph Bafumi Department of Government, Dartmouth College,6108 Silsby HallHanover, NH 03755 e-mail: joseph.bafumi{at}dartmouth.edu Luke Keele Department of Political Science, Ohio State University,2137 Derby Hall, 154 N Oval Mall, Columbus, OH 43210 e-mail: keele.4{at}polisci.osu.edu David Park Department of Political Science, George Washington University,1922 F Street, N.W. 414C, Washington, DC 20052 e-mail: dkp{at}gwu.edu e-mail: bshor{at}uchicago.edu (corresponding author) The analysis of time-series cross-sectional (TSCS) data hasbecome increasingly popular in political science. Meanwhile,political scientists are also becoming more interested in theuse of multilevel models (MLM). However, little work existsto understand the benefits of multilevel modeling when appliedto TSCS data. We employ Monte Carlo simulations to benchmarkthe performance of a Bayesian multilevel model for TSCS data.We find that the MLM performs as well or better than other commonestimators for such data. Most importantly, the MLM is moregeneral and offers researchers additional advantages. Authors' note: A previous version of this article was presentedat the 2005 Midwest Political Science Meeting. We would liketo thank the following for comments and advice in writing thispaper: Andrew Gelman, Nathaniel Beck, Greg Wawro, Sam Cooke,John Londregan, David Brandt. Any errors are our own.     

Protecting Medicare: the best defense is a good offense

Traditional Medicare is being threatened from two political directions. The current Republican coalition, on the right, simply dislikes social insurance in principle. It seeks privatization for its own sake. Another perspective, centrist and well established among political and economic elites, worries that the program is "unaffordable," whatever its basic merits. Defenders of traditional Medicare need to address both threats by explaining why the budgetary fears are misconceived and why privatization is simply a bad idea. In order to do this, they need to take the budgetary high ground, argue more strenuously for short-term cost controls, and criticize the extra spending that the Bush administration has used to encourage private plans within Medicare. Defenders of social insurance should also seek good policy and political allies by proposing that Medicare's network of providers, prices, and administration be made available to employers (and other pools) in much the way that self-insured groups currently rent networks from private insurers.     

Effects of stereotypes and suggestion on memory

In this study, the interactive effect of stereotype and suggestion on accuracy of memory was examined by presenting 645 participants (native Israelis and immigrants from the former Soviet Union and Ethiopia) with three versions of a story about a worker who is waiting in a manager's office for a meeting. All versions were identical except for the worker's name, which implied a Russian or an Ethiopian immigrant or a person of no ethnic origin. Each participant was presented with one version of the story. After an hour delay, the participants' memories were tested via two questionnaires that differed in terms of level of suggestion. Data analyses show that (a) when a suggestion matched the participant's stereotypical perception, the suggestion was incorporated into memory but (b) when the suggestion contradicted the stereotype, it did not influence memory. The conclusion was that recall is influenced by stereotypes but can be enhanced by compatible suggestions.     

Citizenship status and arrest patterns in the United States: evidence from the arrestee drug abuse monitoring program

The purpose of the study was to estimate associations between citizenship status and arrest for crimes among male arrestees. The primary hypothesis was that citizenship status (a rough proxy for immigration) has significant effects on arrest for violent personal crimes, property crimes, and four other selected offenses. Data were derived from the Arrestee Drug Abuse Monitoring (ADAM) Program for the years 2000 through 2002 inclusive. Our sample comprised male arrestees only. Logistic regression models were fitted to the data to estimate the citizenship status-crime relationship. Results showed no significant association between citizenship status and arrest for violent crimes. Non-citizens were 15% less likely than citizens to be arrested for property crimes; they were also less likely to be arrested for weapons offenses and drug offenses. Non-citizens were much less likely to test positive on NIDA-5 drugs than citizens. Non-citizens were, however, 50% more likely than citizens to be arrested for forgery/counterfeiting. It was concluded that public perceptions about the relationship of citizenship status to criminal behavior may be exaggerated and may not be borne out by empirical evidence. Limitations of the study are pointed out, including the fact that in the ADAM data, naturalized immigrants are lumped together with native born citizens.     

Intraobserver Error in Macromorphoscopic Trait Data, ,

As part of a much larger investigation into the use of macromorphoscopic trait data by forensic anthropologists to estimate ancestry from unidentified skeletal remains, we conducted a fourteen‐year (2002–2016) intraobserver error study. Motivated by the development of a large macromorphoscopic database―which will potentially utilize data collected in 2002―quantification of observer error, the impact of technological improvements in macromorphoscopic trait data collection and observer experience is necessary. To maximize comparisons between the two samples, ten macromorphoscopic traits were assessed. Results revealed three patterns of error relating to observer experience, the introduction of new technologies, and error inherent in the method. Overall, this study found the effect of error on macromorphoscopic trait analysis could be predicted and did not significantly impact their utility.     

Diagnosing Crime and Diagnosing Disease‐II: Visual Pattern Perception and Diagnostic Accuracy

Previously, we reviewed how general cognitive processes might be susceptible to bias across both forensic and clinical fields, and how interdisciplinary comparisons could reduce error. We discuss several examples of clinical tasks which are heavily dependent on visual processing, comparing them to eyewitness identification (EI). We review the “constructive” nature of visual processing, and how contextual factors influence both medical experts and witnesses in decision making and recall. Overall, studies suggest common cognitive factors uniting these visual tasks, in both their strengths and shortcomings. Recently forensic sciences have advocated reducing errors by identifying and controlling nonrelevant information. Such efforts could effectively assist medical diagnosis. We suggest potential remedies for cognitive bias in these tasks. These can generalize across the clinical and forensic domains, including controlling the sequencing of contextual factors. One solution is an agnostic primary reading before incorporation of a complete history and interpretation.     

The Future of FOIA: Course Corrections for the Digital Age

After the fiftieth anniversary of the Freedom of Information Act in 2016, criticism of the law is considerable. FOIA has failed to live up to its initial promise of peeling back the layers that too often shroud the federal government in secrecy, frustrating the news media’s efforts to contribute to an informed electorate. Issues with response rates, unorganized systems and subjective interpretations of the act’s exemptions are not uncommon. This article utilizes data gathered from 2008 to 2015 that indicate that across multiple metrics, FOIA has increasingly failed to provide records to requesting parties. The trends suggest that significant overhaul is necessary. Rather than prescribing more amendments that are little more than Band-Aids on a withering dinosaur, this article concludes with a detailed set of recommendations – highlighted by a crowd-sourced request database – that move far from FOIA’s original paper-based model that still rests at its analog core.     

An approach to minimizing legal and reputational risk in Red Team hacking exercises

Robust cyber-resilience depends on sound technical controls and testing of those controls in combination with rigorous cyber-security policies and practices. Increasingly, corporations and other organizations are seeking to test all of these, using methods more sophisticated than mere network penetration testing or other technical audit operations. More sophisticated organizations are also conducting so-called “Red Team” exercises, in which the organization tasks a small team of highly skilled and trained individuals to try to gain unauthorized access to physical and logical company assets and information. While such operations can have real value, they must be planned and conducted with great care in order to avoid violating the law or creating undue risk and reputational harm to the organization. This article explores these sometimes tricky issues, and offers practical risk-based guidance for organizations contemplating these types of exercises.