Abstract not found

Summary. Multicity time series studies of particulate matter and mortality and morbidity have provided evidence that daily variation in air pollution levels is associated with daily variation in mortality counts.These findings served as key epidemiological evidence for the recent review of the US national ambient air quality standards for particulate matter. As a result, methodological issues concerning time series analysis of the relationship between air pollution and health have attracted the attention of the scientific community and critics have raised concerns about the adequacy of current model formulations. Time series data on pollution and mortality are generally analysed by using log-linear, Poisson regression models for overdispersed counts with the daily number of deaths as outcome, the (possibly lagged) daily level of pollution as a linear predictor and smooth functions of weather variables and calendar time used to adjust for timevarying confounders. Investigators around the world have used different approaches to adjust for confounding, making it difficult to compare results across studies. To date, the statistical properties of these different approaches have not been comprehensively compared.To address these issues, we quantify and characterize model uncertainty and model choice in adjusting for seasonal and long-term trends in time series models of air pollution and mortality. First, we

submissions due April 15, 2001. Abstracts submitted after April 15 but before July 15 will be considered but are not guaranteed to be accepted, depending on availability of conference space for sessions. Early registration closes May 15, 2001 to receive early registration discount. The STUDENT PAPER COMPETITION committee will accept student papers until July 15, 2001. This competition includes a cash award of US$500 for the winner. Papers will be judged during the conference and winner announced at the end of the conference. Please encourage students to submit papers following the instructions on the conference web site or by email to TIES.2001@epa.gov. Dr. Agnes M. Herzberg, Queen's University, Canada, will be the TIES President's Lecturer. Dr. Herzberg has organized a series of conferences entitled "Conference on Statistics, Science and Public Policy", with various themes; the latest (April, 2001) was Science and Responsibility. She has been active in the Statistical Society...

Abstract word count: 243

It is routinely asserted that urban air pollution is a major cause of acute respiratory conditions, leading to thousands of hospitalizations each year. The claim is based on inferences from partial correlations between ambient air pollution levels and hospitalization rates. Yet questions persist about the statistical robustness of the epidemiological findings, and controlled experiments have not confirmed the statistical findings. In this paper we present and analyze a new monthly data base showing concentrations of five major air contaminants in 11 large Canadian cities from 1974 to 1994, matched with monthly hospital admission rates by age group for all lung diagnostic categories; as well as a comprehensive set of socioeconomic and meteorological covariates. We compare two estimation approaches: model selection and Bayesian model averaging. Almost all of our estimates of the health effects of air pollution are insignificant. Two pollutant types have significantly negative coefficients, indicating, if interpreted in the standard way, that these pollutants are actually beneficial for health. We do not claim this, but we conclude that the perceived statistical relationship between air pollution and health is not robust. *Address to which all correspondence should be sent.

This paper uses model averaging techniques to identify robust predictors of sovereign default episodes on a pooled database for 46 emerging economies over the period 1980–2004. Sovereign default episodes are defined according to Standard & Poor’s or by non-concessional International Monetary Fund loans in excess of 100 percent of the country’s quota. The authors find that, in addition to the level of indebtedness, the quality of policies and institutions is the best predictor of default episodes in emerging market countries with relatively low levels of external debt. For emerging market countries with a higher level of debt, macroeconomic stability plays a robust role in explaining differences in default probabilities. The paper provides evidence that model averaging can improve out-of-sample prediction of sovereign defaults, and draws policy conclusions for the current crisis based on the results.