A Bayesian Hierarchical Model for the Evaluation of a Website

Consider a website and the surfers visiting its pages. A typical issue of interest, for example while monitoring an advertising campaign, concerns whether a specific page has been designed successfully, i.e. is able to attract surfers or address them to other pages within the site. We assume that the surfing behaviour is fully described by the transition probabilities from one page to another, so that a clickstream (sequence of consecutively visited pages) can be viewed as a finite-state-space Markov chain. We then implement a variety of hierarchical prior distributions on the multivariate logits of the transition probabilities and define, for each page, a content effect and a link effect. The former measures the attractiveness of the page due to its contents, while the latter signals its ability to suggest further interesting links within the site. Moreover, we define an additional effect, representing overall page success, which incorporates both effects previously described. Using WinBUGS, we provide estimates and credible intervals for each of the above effects and rank pages accordingly.     

A Non-parametric Frailty Model for Temporally Clustered Multivariate Failure Times

Abstract A model is introduced here for multivariate failure time data arising from heterogenous populations. In particular, we consider a situation in which the failure times of individual subjects are often temporally clustered, so that many failures occur during a relatively short age interval. The clustering is modelled by assuming that the subjects can be divided into ‘internally homogenous’ latent classes, each such class being then described by a time‐dependent frailty profile function. As an example, we reanalysed the dental caries data presented earlier in Härkänen et al. [Scand. J. Statist. 27 (2000) 577], as it turned out that our earlier model could not adequately describe the observed clustering.     

‘I hereby come out’: What sort of speech act is coming out?

The process of coming out linguistically as a lesbian or as a gay man is occasionally referred to as a speech act (e.g. Harvey 1997: 72; Liang 1997: 293). This analogy is taken as a starting point to explore the extent to which coming out is a performative act, and what sort of speech act coming out may be. The discussion draws on the perspective of both the speaker and the hearer to consider how the acts involved in coming out are open to interpretation. Parallels are drawn between the act of coming out as a lesbian or a gay man and other instances of self–disclosure or of individuals' constructions of new facets of their identities.     

Applicability of a Neuroprobabilistic Integral Risk Index for the Environmental Management of Polluted Areas: A Case Study

Recently, we developed a GIS-Integrated Integral Risk Index (IRI) to assess human health risks in areas with presence of environmental pollutants. Contaminants were previously ranked by applying a self-organizing map (SOM) to their characteristics of persistence, bioaccumulation, and toxicity in order to obtain the Hazard Index (HI). In the present study, the original IRI was substantially improved by allowing the entrance of probabilistic data. A neuroprobabilistic HI was developed by combining SOM and Monte Carlo analysis. In general terms, the deterministic and probabilistic HIs followed a similar pattern: polychlorinated biphenyls (PCBs) and light polycyclic aromatic hydrocarbons (PAHs) were the pollutants showing the highest and lowest values of HI, respectively. However, the bioaccumulation value of heavy metals notably increased after considering a probability density function to explain the bioaccumulation factor. To check its applicability, a case study was investigated. The probabilistic integral risk was calculated in the chemical/petrochemical industrial area of Tarragona (Catalonia, Spain), where an environmental program has been carried out since 2002. The risk change between 2002 and 2005 was evaluated on the basis of probabilistic data of the levels of various pollutants in soils. The results indicated that the risk of the chemicals under study did not follow a homogeneous tendency. However, the current levels of pollution do not mean a relevant source of health risks for the local population. Moreover, the neuroprobabilistic HI seems to be an adequate tool to be taken into account in risk assessment processes.     

Using Probabilistic Terrorism Risk Modeling for Regulatory Benefit-Cost Analysis: Application to the Western Hemisphere Travel Initiative in the Land Environment

This article presents a framework for using probabilistic terrorism risk modeling in regulatory analysis. We demonstrate the framework with an example application involving a regulation under consideration, the Western Hemisphere Travel Initiative for the Land Environment, (WHTI‐L). First, we estimate annualized loss from terrorist attacks with the Risk Management Solutions (RMS) Probabilistic Terrorism Model. We then estimate the critical risk reduction, which is the risk‐reducing effectiveness of WHTI‐L needed for its benefit, in terms of reduced terrorism loss in the United States, to exceed its cost. Our analysis indicates that the critical risk reduction depends strongly not only on uncertainties in the terrorism risk level, but also on uncertainty in the cost of regulation and how casualties are monetized. For a terrorism risk level based on the RMS standard risk estimate, the baseline regulatory cost estimate for WHTI‐L, and a range of casualty cost estimates based on the willingness‐to‐pay approach, our estimate for the expected annualized loss from terrorism ranges from $2.7 billion to $5.2 billion. For this range in annualized loss, the critical risk reduction for WHTI‐L ranges from 7% to 13%. Basing results on a lower risk level that results in halving the annualized terrorism loss would double the critical risk reduction (14–26%), and basing the results on a higher risk level that results in a doubling of the annualized terrorism loss would cut the critical risk reduction in half (3.5–6.6%). Ideally, decisions about terrorism security regulations and policies would be informed by true benefit‐cost analyses in which the estimated benefits are compared to costs. Such analyses for terrorism security efforts face substantial impediments stemming from the great uncertainty in the terrorist threat and the very low recurrence interval for large attacks. Several approaches can be used to estimate how a terrorism security program or regulation reduces the distribution of risks it is intended to manage. But, continued research to develop additional tools and data is necessary to support application of these approaches. These include refinement of models and simulations, engagement of subject matter experts, implementation of program evaluation, and estimating the costs of casualties from terrorism events.     

Urban Power Structures and Publicly Financed Stadiums

Urban elites are increasingly addressing local social problems though policies that turn their cities into tourist destinations. Often at the heart of these policies are new publicly financed sports stadiums. Ironically, this strategy is flourishing despite near-unanimous academic criticism, and increasing public skepticism, about this approach. Our research addresses this contradiction by exploring how and why powerful decisionmakers continue supporting publicly financed stadiums. We rely on local growth coalition theory to explore this topic because it offers analytical advantages, including looking beyond local sports teams as the focal point of these initiatives, addressing the variation in the outcomes of these initiatives, and acknowledging that policymakers are predisposed toward supporting these initiatives but that this predisposition does not always result in success.     

Risk,trust and knowledge networks in farmers’ learning

This study reports on New Zealand dairy farmers’ access to and use of information as mediated through conditions of risk and trust within the context of their interpersonal social networks. We located participants’ reports of their information use within their perceived environments of trust and risk, following Giddens's [1990. The consequences of modernity. Polity Press, Stanford, CA] typology of trust and risk in pre-modernity and modernity. The research participants were constant users of interpersonal and print information from numerous sources, and monitored their incoming data in the light of strategic needs, reflecting their roles as both farming practitioners and business owners. Socio-spatial knowledge networks (SSKNs) combine individuals’ explanatory cognitive models of information acquisition and use with a micro-geographical analysis of their interpersonal networks. The participants showed characteristics of pre-modern, modern and even post-modern society in respect of their use of complex interactional forms, as well as a blending of individualistic and communitarian practices and concerns in their professional and personal lives.     

Which Wald statistic? Choosing a parameterization of the Wald statistic to maximize power in k-sample generalized estimating equations

The Wald statistic is known to vary under reparameterization. This raises the question: which parameterization should be chosen, in order to optimize power of the Wald statistic? We specifically consider k-sample tests of generalized linear models (GLMs) and generalized estimating equations (GEEs) in which the alternative hypothesis contains only two parameters. An example is presented in which such an alternative hypothesis is of interest. Amongst a general class of parameterizations, we find the parameterization that maximizes power via analysis of the non-centrality parameter, and show how the effect on power of reparameterization depends on sampling design and the differences in variance across samples. There is no single parameterization with optimal power across all alternatives. The Wald statistic commonly used under the canonical parameterization is optimal in some instances but it performs very poorly in others. We demonstrate results by example and by simulation, and describe their implications for likelihood ratio statistics and score statistics. We conclude that due to poor power properties, the routine use of score statistics and Wald statistics under the canonical parameterization for GEEs is a questionable practice.     

Bootstrap Estimation of Benchmark Doses and Confidence Limits with Clustered Quantal Data

The benchmark dose (BMD) is an exposure level that would induce a small risk increase (BMR level) above the background. The BMD approach to deriving a reference dose for risk assessment of noncancer effects is advantageous in that the estimate of BMD is not restricted to experimental doses and utilizes most available dose-response information. To quantify statistical uncertainty of a BMD estimate, we often calculate and report its lower confidence limit (i.e., BMDL), and may even consider it as a more conservative alternative to BMD itself. Computation of BMDL may involve normal confidence limits to BMD in conjunction with the delta method. Therefore, factors, such as small sample size and nonlinearity in model parameters, can affect the performance of the delta method BMDL, and alternative methods are useful. In this article, we propose a bootstrap method to estimate BMDL utilizing a scheme that consists of a resampling of residuals after model fitting and a one-step formula for parameter estimation. We illustrate the method with clustered binary data from developmental toxicity experiments. Our analysis shows that with moderately elevated dose-response data, the distribution of BMD estimator tends to be left-skewed and bootstrap BMDL s are smaller than the delta method BMDL s on average, hence quantifying risk more conservatively. Statistically, the bootstrap BMDL quantifies the uncertainty of the true BMD more honestly than the delta method BMDL as its coverage probability is closer to the nominal level than that of delta method BMDL. We find that BMD and BMDL estimates are generally insensitive to model choices provided that the models fit the data comparably well near the region of BMD. Our analysis also suggests that, in the presence of a significant and moderately strong dose-response relationship, the developmental toxicity experiments under the standard protocol support dose-response assessment at 5% BMR for BMD and 95% confidence level for BMDL.