Eliciting prior information to enhance the predictive performance of bayesian graphical models

Both knowledge-based systems and statistical models are typically concerned with making predictions about future observables. Here we focus on assessment of predictive performance and provide two techniques for improving the predictive performance of Bayesian graphical models. First, we present Bayesian model averaging, a technique for accounting for model uncertainty.

Second, we describe a technique for eliciting a prior distribution for competing models from domain experts. We explore the predictive performance of both techniques in the context of a urological diagnostic problem.     

The relationship between transformational leadership behaviors,psychological, and training outcomes in elite military recruits

Two studies examined the effects of a differentiated model of transformational leadership on follower outcomes. In Study 1, 484 UK Royal Marine recruits completed questionnaires about their trainers' leadership behaviors and their own attitudes towards training. Training outcome was measured as successful completion of training or non-completion. Discriminant function analyses identified that fostering acceptance of group goals, inspirational motivation, appropriate role model, individual consideration, and contingent reward significantly discriminated between pass and failure. A separate discriminant function analyses revealed that the attitudinal variables of self-confidence, resilience, and satisfaction also successfully discriminated between pass and failure. Study 2 used a true experimental design to examine the effectiveness of a transformational leadership intervention. Participants were 85 experimental and 67 control recruits who completed questionnaires at weeks 5 and 15 of recruit training. Results revealed that 3 of the 5 key leadership behaviors, and all of the 3 recruit attitudinal variables measured, were significantly enhanced by the intervention.     

Using Single Respondents in Strategy Research

A survey of empirical papers has revealed a number of studies that have used single respondents, usually the CEO or a top management team member, to make inferences about certain strategic dimensions of their business. In this note, we highlight some of the problems in using this approach. We base our comments on a study of 76 businesses where multiple respondents were accessed. We conclude that, because of often wide variations in responses between members of the same top team (measured by Cronbach alpha), the use of single respondents in this type of strategy research is probably unreliable.     

Changes in the association between job decision latitude and work engagement at different levels of work experience: A 10-year longitudinal study

ABSTRACT

This study set out to explore the effects of accumulating work experience on the association between job decision latitude and its interaction with job demands and work engagement. Our ten-year longitudinal study followed 333 junior physicians in postgraduate training at baseline. We used self-report measures in four assessment waves, and we conducted path analyses to investigate linear and curvilinear regression effects. Results show that high job decision latitude was associated with high work engagement at all levels of work experience, with strongest associations at baseline and after ten years. Only for novices did job decision latitude buffer the negative association between job demands and work engagement. At the stage of high work experience, low levels of job decision latitude were weakly associated with work engagement, whereas with higher levels of job decision latitude, the positive association seemed to strengthen. Our findings indicate that job decision latitude is a key job resource at all stages of work experience, with stronger effects among novices and experts. Organisations’ work design efforts should include job decision latitude to promote work engagement across employees’ different career stages, with consideration to job entrants and experts in order to tailor specific work design solutions.     

Model comparison for a complex ecological system

Summary.  The ecological consequences of climate change and its interaction with other environmental pressures, such as nutrient pollution, are little understood. For freshwater ecosystems, knowledge of these combined effects is required for water resource management and in particular for successful implementation of the European Community Water Framework Directive 2000, which requires that all surface-waters should be at, or above, 'good status' by 2016. Statistical analysis of detailed long-term environmental data sets can be used to explore these combined effects. Loch Leven (Scotland) has been routinely monitored since 1967, providing one of the largest and most extensive of such data sets. Over this period there has been evidence of climate change and a period of eutrophication and recovery at the loch. Transfer functions, additive models and varying-coefficient models are used to explore the complex ecological system at Loch Leven with the aim of obtaining insight into the combined effects of climate change and eutrophication on water quality.     

What are dynamic capabilities and are they a useful construct in strategic management?

The dynamic capability perspective extends the resource-based view argument by addressing how valuable, rare, difficult to imitate and imperfectly substitutable resources can be created and how the current stock of valuable resources can be refreshed in changing environments. The concept of dynamic capabilities emerged in the 1990s, and the field has advanced considerably since. This paper presents a review as well as a synthesis of the extant literature. This synthesis first highlights, that dynamic capabilities are shaped by enabling and inhibiting variables within and outside the firm, including the perceptions and motivations of managers; secondly, it identifies processes that create dynamic capabilities; and thirdly, it explains that dynamic capabilities do not automatically lead to performance improvements. Finally, the paper addresses some areas of confusion and contradiction that hamper the development of the literature.