A Semiparametric Bayesian Model for Circular-Linear Regression

Circular data are observations that are represented as points on a unit circle. Times of day and directions of wind are two such examples. In this work, we present a Bayesian approach to regress a circular variable on a linear predictor. The regression coefficients are assumed to have a nonparametric distribution with a Dirichlet process prior. The semiparametric Bayesian approach gives added flexibility to the model and is useful especially when the likelihood surface is ill behaved. Markov chain Monte Carlo techniques are used to fit the proposed model and to generate predictions. The method is illustrated using an environmental data set.     

Seroadaptive Strategies of Vancouver Gay and Bisexual Men in a Treatment as Prevention Environment

British Columbia’s treatment as prevention policy has provided free access to highly active antiretroviral therapy (HAART) to all HIV-positive provincial residents since 1996. One outcome is an increase in HIV-positive gay and bisexual men (GBM) with suppressed viral loads. Previous cross-sectional analyses indicated that some Vancouver GBM now recognize condomless anal sex with men on HAART who report a suppressed viral load as a seroadaptive strategy. To test the hypothesis that this new strategy, termed viral load sorting (VLS), is recognized and used among by GBM in the Momentum Health Study, we analyzed longitudinal data for HIV-negative/unknown (n = 556) and HIV-positive (n = 218) serostatus participants. Analyses indicated that both groups reported VLS, and that serostatus and Treatment Optimism Scale scores were significant determinants in frequency and use. Results exemplify the medicalization of sex and Rogers’ Diffusion Of Preventative Innovations Model, and they have important implications for HIV research and GBM sexual decision-making.     

Two-stage approaches to the analysis of occupancy data I: the homogeneous case (analysis of occupancy data)

Abstract

Occupancy models are used in statistical ecology to estimate species dispersion. The two components of an occupancy model are the detection and occupancy probabilities, with the main interest being in the occupancy probabilities. We show that for the homogeneous occupancy model there is an orthogonal transformation of the parameters that gives a natural two-stage inference procedure based on a conditional likelihood. We then extend this to a partial likelihood that gives explicit estimators of the model parameters. By allowing the separate modeling of the detection and occupancy probabilities, the extension of the two-stage approach to more general models has the potential to simplify the computational routines used there.     

CRIMINALS' RESPONSE TO CHANGING CRIME LUCRE

How do criminals respond to changes in the benefit from committing a successful crime? This question is relevant for understanding the effectiveness of crime‐fighting policies that reduce demand for illegal goods, disrupt black markets, and otherwise eliminate cheaper avenues to illicit gain. However, the literature has not sufficiently addressed this question, partly because finding a reliable measure of crime lucre is difficult. Using proprietary data on cargo theft, I match historical prices of various goods with their thefts and estimate the price elasticity of theft to be 1.225 over a cumulative 7‐month horizon. (JEL K42)     

BANKING AND PUBLIC POLICY: TOO BIG TO FAIL

“Too big to fail” (TBTF) is a major policy issue in banking. Large bank failures may impose losses on depositors and creditors that may impose large collateral damage on other financial institutions and beyond. Regulators are frequently incentivized either to delay recognizing a bank's insolvency or fail the bank but protect its creditors against loss. This paper argues that, while there is wide agreement that the cost of protecting creditors in the resolution of large financial institutions is excessively high, it is difficult to prevent this practice for a number of reasons. Until these disagreements are settled, TBTF will survive. (JEL G01, G18, G28)     

Principles for Conduct of Pest Risk Analyses: Report of an Expert Workshop

The North American Free Trade Agreement (NAFTA) and the General Agreement on Tariffs and Trade (GATT) have focused attention on risk assessment of potential insect, weed, and animal pests and diseases of livestock. These risks have traditionally been addressed through quarantine protocols ranging from limits on the geographical areas from which a product may originate, postharvest disinfestation procedures like fumigation, and inspections at points of export and import, to outright bans. To ensure that plant and animal protection measures are not used as nontariff trade barriers, GATT and NAFTA require pest risk analysis (PRA) to support quarantine decisions. The increased emphasis on PRA has spurred multiple efforts at the national and international level to design frameworks for the conduct of these analyses. As approaches to pest risk analysis proliferate, and the importance of the analyses grows, concerns have arisen about the scientific and technical conduct of pest risk analysis. In January of 1997, the Harvard Center for Risk Analysis (HCRA) held an invitation-only workshop in Washington, D.C. to bring experts in risk analysis and pest characterization together to develop general principles for pest risk analysis. Workshop participants examined current frameworks for PRA, discussed strengths and weaknesses of the approaches, and formulated principles, based on years of experience with risk analysis in other setting and knowledge of the issues specific to analysis of pests. The principles developed highlight the both the similarities of pest risk analysis to other forms of risk analysis, and its unique attributes.     

The programmable society and the individual as a unit of data

The authors commence this paper with discussion of an extract from H.G. Wells' novel of 1920. They use this extract to remind the reader that social innovation is not a new problem and that imaginative writers may be a source of guidance on such matters. Specifically they deal with the social options created by ‘computer-driven’ information systems. They argue that this aspect of the problem of science and people is one which has a great and seriously under-estimated potential to influence the quality of life in a future society.