Models, assumptions and model checking in ecological regressions

Ecological regression is based on assumptions that are untestable from aggregate data. However, these assumptions seem more questionable in some applications than in others. There has been some research on implicit models of individual data underlying aggregate ecological regression modelling. We discuss ways in which these implicit models can be checked from aggregate data. We also explore the differences in applications of ecological regressions in two examples: estimating the effect of radon on lung cancer in the United States and estimating voting patterns for different ethnic groups in New York City.     

States of nature and states of mind: a generalized theory of decision-making

Canonical economic agents act so as to maximize a single, representative, utility function. However, there is accumulating evidence that heterogeneity in thought processes may be an important determinant of individual behavior. This paper investigates the implications of a vector-valued generalization of the Expected Utility paradigm, which permits agents either to deliberate as per Homo economics, or to act impulsively. This generalized decision theory is applied to explain the crowding-out effect, irrational educational investment decisions, persistent social inequalities, the pervasive influence of non-cognitive ability on socio-economic outcomes, and the dynamic relationships between non-cognitive ability, cognitive ability, and behavioral biases. These results suggest that the generalized decision theory warrants further investigation.

     

The problem of suffering as a driving force of rationalization and social change

This article documents and analyses a reconstructed Weberian conception of the problem of suffering. In this setting a focus is brought to how the problem of suffering is constituted in the dynamic interplay between, on the one hand, the compulsion to impose rational sense and order on the world, and on the other, the necessity to find a means to satiate charismatic needs. The discussion highlights Weber's account of the tendency for problems of suffering to increase in volume and scale along with the intensification and spread of modern processes of rationalization. It offers a case for the development of further sociological inquiries into the role played by experiences of the problem of suffering within the dynamics of social and cultural change.     

Using innovation to address HIV, AIDS, and environment links: intervention case studies from Zimbabwe, Uganda, and Malawi

This article presents three cross-cutting intervention case studies that address HIV, AIDS, and natural resources management in an integrated manner through innovative programming in Zimbabwe, Uganda, and Malawi. In Zimbabwe, a religious-based non-governmental group and two community organizations work together to build skills of HIV- and AIDS-vulnerable children in rural areas to meet dietary and income needs, while using natural resources sustainably. In Uganda, various government agencies and NGO actors work together to improve the food security of HIV-affected households at the national, district, sub-district, and village levels. Finally, in Malawi, a conservation organization incorporates HIV and AIDS awareness and programming into its operations and projects. Each case study presents pioneering approaches to simultaneously addressing the pressures on conservation initiatives, food security/agricultural production, income generation/livelihoods, and social and health care systems. They also provide lessons for expanding interventions and partnerships.     

Hope over fear: social work education towards 2025

ABSTRACT

Prediction of possible futures is fraught with dangers. Neither the global economic crisis which erupted in 2008 nor the political earthquake which shook Scotland over the issue of independence during 2014 was foreseen by many commentators, if indeed any. Given these experiences, predicting where social work education might be in 2025 is a potentially hazardous enterprise. Nevertheless, the recent resurgence of interest in utopian thinking reflects a widely felt desire to go beyond ‘capitalist realism’ and to envisage different possibilities – a desire also reflected in political developments in Greece and Spain. This development is primarily in reaction to the dominance of another form of utopian (or dystopian) thinking: neo-liberalism, with its message that ‘there is no alternative’. In this paper, I will argue that that search for alternatives has important implications for social work and social work education. Following a discussion of the ways in which neo-liberalism has shaped the profession over two decades, the paper will identify current challenges to neo-liberal social work and social work education and more widely, to the politics of austerity. Drawing on examples from different countries, I will argue that this ‘new radicalism’ points the way to a more politically engaged social work education.     

Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC

Leave-one-out cross-validation (LOO) and the widely applicable information criterion (WAIC) are methods for estimating pointwise out-of-sample prediction accuracy from a fitted Bayesian model using the log-likelihood evaluated at the posterior simulations of the parameter values. LOO and WAIC have various advantages over simpler estimates of predictive error such as AIC and DIC but are less used in practice because they involve additional computational steps. Here we lay out fast and stable computations for LOO and WAIC that can be performed using existing simulation draws. We introduce an efficient computation of LOO using Pareto-smoothed importance sampling (PSIS), a new procedure for regularizing importance weights. Although WAIC is asymptotically equal to LOO, we demonstrate that PSIS-LOO is more robust in the finite case with weak priors or influential observations. As a byproduct of our calculations, we also obtain approximate standard errors for estimated predictive errors and for comparison of predictive errors between two models. We implement the computations in an R package called loo and demonstrate using models fit with the Bayesian inference package Stan.