Statistical Control for Road Pavements

Roads are assets vital to the economies of nations, particularly those with such low population density as Australia and New Zealand. The quality of road construction is of great importance. The application of statistical techniques to determining whether the construction of road pavement is being carried out to the design specification can be of great benefit to administrators and contractors. In this paper, a method is presented for obtaining control limits for a means chart when correlated observations are taken over a grid on a two‐dimensional surface. Data from a closely monitored road construction project are used to determine if the construction process can be considered ‘in control’. The nature of the correlation structure is determined using ARIMA models. The ultimate aim of the control charts and specifications is to control base course thickness. Controlling the pavement surfaces is a means to this end.     

Measuring Research Performance in Business and Management Studies in the United Kingdom: The 1992 Research Assessment Exercise*

This paper presents the results of a statistical analysis of the 1992 Research Assessment Exercise (RAE92) for the 85 business and management departments in UK institutions of higher education. The aims of the paper are, first, to construct a set of quantitative indicators of research outputs and inputs which can be used by individual departments for comparative purposes and second, to investigate the extent to which variations between departments in the research rating obtained in the RAE92 can be explained by these indicators of research outputs and inputs. One of the main findings is that around 80 per cent of the variation in the research rating obtained by business and management departments are accurately predicted by a regression model that contains only a small number of regressors. These include: size of department, articles in refereed academic journals, research students and research income.     

Dependency of Crop Production between Global Breadbaskets: A Copula Approach for the Assessment of Global and Regional Risk Pools

As recent events have shown, simultaneous crop losses in different parts of the world can cause serious risks to global food security. However, to date, little is known about the spatial dependency of lower than expected crop yields from global breadbaskets. This especially applies in the case of extreme events, i.e., where one or more breadbaskets are experiencing far below average yields. Without such information, risk management approaches cannot be applied and vulnerability to extremes may remain high or even increase in the future around the world. We tackle both issues from an empirical perspective focusing on wheat yield. Interdependencies between historically observed wheat yield deviations in five breadbaskets (United States, Argentina, India, China, and Australia) are estimated via copula approaches that can incorporate increasing tail dependencies. In doing so, we are able to attach probabilities to interregional as well as global yield losses. To address the robustness of our results, we apply three different methods for constructing multivariate copulas: vine copulas, ordered coupling using a minimax approach, and hierarchical structuring. We found interdependencies between states within breadbaskets that led us to the conclusion that risk pooling for extremes is less favorable on the regional level. However, notwithstanding evidence of global climatic teleconnections that may influence crop production, we also demonstrate empirically that wheat production losses are independent between global breadbaskets, which strengthens the case for interregional risk pooling strategies. We argue that through interregional risk pooling, postdisaster liabilities of governments and international donors could be decreased.     

Gambling and problem gambling among clients,and staff attitudes,in an alcohol and drug problems treatment service in the English Midlands

A self-completion questionnaire containing items about recent gambling and problem gambling used in the British Gambling Prevalence Survey (BGPS) was completed by 274 clients of alcohol and drug problems services in the English West Midlands. Compared to BGPS data, matched to the client sample for age and sex, clients were significantly more likely in the previous 12 months to have engaged in forms of gambling carrying a high risk of problems, to have engaged in four or more separate forms of gambling, and to have scores on a brief form of the South Oaks Gambling Screen (SOGS) indicating problem gambling. A self-completion scale on attitudes towards working with problem gamblers, with good psychometric properties according to present data, was completed by 91 members of staff of the same services. With the exception of a sub-scale measuring perceived adequacy of knowledge and skills for working with problem gamblers, staff expressed moderately positive attitudes. It is concluded that British substance misuse services are appropriate locations for the development of services for problem gambling.     

BAYESIAN HYPER‐LASSOS WITH NON‐CONVEX PENALIZATION

The Lasso has sparked interest in the use of penalization of the log‐likelihood for variable selection, as well as for shrinkage. We are particularly interested in the more‐variables‐than‐observations case of characteristic importance for modern data. The Bayesian interpretation of the Lasso as the maximum a posteriori estimate of the regression coefficients, which have been given independent, double exponential prior distributions, is adopted. Generalizing this prior provides a family of hyper‐Lasso penalty functions, which includes the quasi‐Cauchy distribution of Johnstone and Silverman as a special case. The properties of this approach, including the oracle property, are explored, and an EM algorithm for inference in regression problems is described. The posterior is multi‐modal, and we suggest a strategy of using a set of perfectly fitting random starting values to explore modes in different regions of the parameter space. Simulations show that our procedure provides significant improvements on a range of established procedures, and we provide an example from chemometrics.