Statistical methods to evaluate health effects associated with major sources of air pollution: a case-study of breathing patterns during exposure to concentrated Boston air particles

Summary.  We conduct a case-study evaluating the source-specific effects of particulate matter on respiratory function. Using a structural equation approach, we assess the effect of different receptor models on the estimated source-specific effects for univariate respiratory response. Furthermore, we extend the structural equation model by placing a factor analysis model on the response to represent the measured respiratory responses in terms of underlying respiratory patterns. We estimate the particulate matter source-specific effects on respiratory rate, accentuated normal breathing and airway irritation and find a strong increase in airway irritation that is associated with exposure to motor vehicle particulate matter.     

Bayesian inference for Poisson and multinomial log-linear models

Categorical data frequently arise in applications in the Social Sciences. In such applications, the class of log-linear models, based on either a Poisson or (product) multinomial response distribution, is a flexible model class for inference and prediction. In this paper we consider the Bayesian analysis of both Poisson and multinomial log-linear models. It is often convenient to model multinomial or product multinomial data as observations of independent Poisson variables. For multinomial data, Lindley (1964) [20] showed that this approach leads to valid Bayesian posterior inferences when the prior density for the Poisson cell means factorises in a particular way. We develop this result to provide a general framework for the analysis of multinomial or product multinomial data using a Poisson log-linear model. Valid finite population inferences are also available, which can be particularly important in modelling social data. We then focus particular attention on multivariate normal prior distributions for the log-linear model parameters. Here, an improper prior distribution for certain Poisson model parameters is required for valid multinomial analysis, and we derive conditions under which the resulting posterior distribution is proper. We also consider the construction of prior distributions across models, and for model parameters, when uncertainty exists about the appropriate form of the model. We present classes of Poisson and multinomial models, invariant under certain natural groups of permutations of the cells. We demonstrate that, if prior belief concerning the model parameters is also invariant, as is the case in a ‘reference’ analysis, then the choice of prior distribution is considerably restricted. The analysis of multivariate categorical data in the form of a contingency table is considered in detail. We illustrate the methods with two examples.     

A Systems Approach to the Policy‐Level Risk Assessment of Exotic Animal Diseases: Network Model and Application to Classical Swine Fever

Exotic animal diseases (EADs) are characterized by their capacity to spread global distances, causing impacts on animal health and welfare with significant economic consequences. We offer a critique of current import risk analysis approaches employed in the EAD field, focusing on their capacity to assess complex systems at a policy level. To address the shortcomings identified, we propose a novel method providing a systematic analysis of the likelihood of a disease incursion, developed by reference to the multibarrier system employed for the United Kingdom. We apply the network model to a policy‐level risk assessment of classical swine fever (CSF), a notifiable animal disease caused by the CSF virus. In doing so, we document and discuss a sequence of analyses that describe system vulnerabilities and reveal the critical control points (CCPs) for intervention, reducing the likelihood of U.K. pig herds being exposed to the CSF virus.     

Cigarette Use Among Young Adults: Comparisons Between 2-Year College Students, 4-Year College Students,and Those Not in College

Abstract

Objective: To examine cigarette smoking among young adults based on education status. Participants: Community-based sample of 2,694 young adults in the United States Methods: The authors compared 3 groups—those not in college with no college degree, 2-year college students/graduates, 4-year college students/graduates—on various smoking measures: ever smoked, smoked in past month, smoked in past week, consider self a smoker, began smoking before age 15, smoked over 100 cigarettes in lifetime, ever tried to quit, and plan to quit in next year. Results: The authors found that for nearly all the smoking measures, the 4-year college group was at lowest risk, the noncollege group was at highest risk, and the 2-year college group represented a midpoint. Differences between groups remain after adjusting for parents’ education and other potential confounding factors. Conclusions: Smoking behaviors clearly differ between the 2-year, 4-year, and no college groups. Interventions should be tailored for each group.     

Communities and academics: A developing dialogue

Intentional communities have been described as laboratories for testing and demonstrating new ideologies and social structures. This paper focuses on the part played by academics in disseminating the results of these ‘social experiments’ to a wider audience. The relationship between intentional communities and academics has often been difficult, with mistrust on both sides. By looking at the underlying needs of both, this paper intends to throw light on the source of some of the difficulties, so that both may better fulfil their important social functions. A case study is reported to illustrate the main points. The paper concludes with advice to both intentional communities and academics for building better relationships and developing co-operative projects.     

A note on glejser's test for heteroscedasticity

Glejser published a test on the residuals of a regression model where the parameters are estimated by OLS that purports to detect “mixed” heteroscedasticity. This note addresses the problem of detecting this type heteroscedasticity from,both a theoretical and pragmatic point of view. We conclude that “mixed” heteroscedasticity cannot be separated from non zero expected errors and thus cannot be detected using Glejser s technique.