Structure and uncertainty: Graphical models for understanding complex data

Statistics is fundamental to making sense of the complexity of modern science. From the micro-level of the human genome to the macro-level of the universe, scientists need statistical models to help them extract meaning from empirical observations. Graphical models have been used across a wide variety of disciplines for building multivariate probabilistic models to represent, and draw inference about, complex phenomena. Nicky Best and Peter Green explain the ideas behind graphical models and show how they can be used to help tackle the challenges of complex statistical problems.     

What might Normalisation Process Theory bring to policy implementation studies? Learning lessons and uncovering questions through a case study of the profound implementation failure of a new policing policy

Normalisation Process Theory (NPT), used nationally and internationally to explore implementation within health services research, is used for the first time within policing to understand profound policy implementation failure and to generate broader discussion of policy implementation theory. The policy in question (Police to Primary Care [P2PC]) was an intervention designed to notify GPs when women are assessed by police as at high risk of future domestic abuse. Designed to improve interagency communication, it took place amidst radical organisational change. Using qualitative interviews with domestic abuse specialist and frontline officers, this paper addresses how NPT helps to explain the (non)implementation of P2PC, how such an analysis differs from other policy implementation approaches, and what this means for our understandings of policy implementation more broadly. NPT proved useful in understanding mechanisms leading to (non)implementation of the intervention: fuzzy alignment with existing practice, faulty communication of purpose, and inattention to discretionary implementation spaces. It helped us understand why the intervention came to be invisible. Dwarfed by its organisational context, made institutionally hard to read by a lack of formal protocols, and given restricted view to police officers, it was compromised by a failure to instigate systems of organisational learning. More broadly, NPT helped reveal practices intersecting top‐down and bottom‐up implementation theory. The paper concludes by asking how NPT and theories of street‐level bureaucracy might be better used in tandem and, particularly, how this might help explorations of policy implementation where human actors are joined by technological actors in interpreting and making policy in vivo.     

An Introduction to ‘Benefit of the Doubt’ Composite Indicators

Despite their increasing use, composite indicators remain controversial. The undesirable dependence of countries’ rankings on the preliminary normalization stage, and the disagreement among experts/stakeholders on the specific weighting scheme used to aggregate sub-indicators, are often invoked to undermine the credibility of composite indicators. Data envelopment analysis may be instrumental in overcoming these limitations. One part of its appeal in the composite indicator context stems from its invariance to measurement units, which entails that a normalization stage can be skipped. Secondly, it fills the informational gap in the ‘right’ set of weights by generating flexible ‘benefit of the doubt’-weights for each evaluated country. The ease of interpretation is a third advantage of the specific model that is the main focus of this paper. In sum, the method may help to neutralize some recurring sources of criticism on composite indicators, allowing one to shift the focus to other, and perhaps more essential stages of their construction. An abridged version of this paper was presented at the Workshop on European Indicators and Scoreboards, organised by DG Education and the Joint Research Centre within the auspices of CRELL, in Brussels, October 24–25, 2005.     

Bidirectionality in kindergarten children's school readiness and emotional regulation

This study examined the longitudinal associations of emotional regulation and school readiness among Chinese kindergarten children. Data were collected from 523 children (mean age at time 1 = 52.42 months; 52.9% male) at two time points separated by approximately 1 year in Hong Kong, China. At times 1 and 2, children's school readiness was assessed by their teachers and parents while their emotional regulation was directly tested. Teachers also rated children's emotional regulation at both time points. The results from the two reciprocal path models showed that, controlling for the corresponding autoregressive effects, school readiness at time 1 was predictive of emotional regulation at time 2. However, emotional regulation at time 1 did not emerge as a significant predictor of school readiness at time 2. Theoretically, these results underscored children's school readiness as a potential contributor to their emotional regulation. Practically, these findings suggested the utility of supporting children's school readiness to foster their emotional regulation.     

Generic reversible jump MCMC using graphical models

Markov chain Monte Carlo techniques have revolutionized the field of Bayesian statistics. Their power is so great that they can even accommodate situations in which the structure of the statistical model itself is uncertain. However, the analysis of such trans-dimensional (TD) models is not easy and available software may lack the flexibility required for dealing with the complexities of real data, often because it does not allow the TD model to be simply part of some bigger model. In this paper we describe a class of widely applicable TD models that can be represented by a generic graphical model, which may be incorporated into arbitrary other graphical structures without significantly affecting the mechanism of inference. We also present a decomposition of the reversible jump algorithm into abstract and problem-specific components, which provides infrastructure for applying the method to all models in the class considered. These developments represent a first step towards a context-free method for implementing TD models that will facilitate their use by applied scientists for the practical exploration of model uncertainty. Our approach makes use of the popular WinBUGS framework as a sampling engine and we illustrate its use via two simple examples in which model uncertainty is a key feature.