Variational Bayes with synthetic likelihood

Synthetic likelihood is an attractive approach to likelihood-free inference when an approximately Gaussian summary statistic for the data, informative for inference about the parameters, is available. The synthetic likelihood method derives an approximate likelihood function from a plug-in normal density estimate for the summary statistic, with plug-in mean and covariance matrix obtained by Monte Carlo simulation from the model. In this article, we develop alternatives to Markov chain Monte Carlo implementations of Bayesian synthetic likelihoods with reduced computational overheads. Our approach uses stochastic gradient variational inference methods for posterior approximation in the synthetic likelihood context, employing unbiased estimates of the log likelihood. We compare the new method with a related likelihood-free variational inference technique in the literature, while at the same time improving the implementation of that approach in a number of ways. These new algorithms are feasible to implement in situations which are challenging for conventional approximate Bayesian computation methods, in terms of the dimensionality of the parameter and summary statistic.     

Fall incidence in a state psychiatric hospital in Singapore

The results of a retrospective study of fall incidence during a 1-year period in a psychiatric hospital in Singapore are reported, involving 309 patients who fell one or more times during their stay. The profile of inpatients involved in falls was identified. Data were derived from standard incident forms completed whenever patients had a fall. A total of 348 falls were identified for the 1-year period. Results revealed higher fall rates in younger male epileptic patients; elderly female patients, age 70 and older with depression or dementia; individuals with concomitant medical conditions; and patients taking three types of medication. The majority of falls occurred when the activity level was high in the ward. Young epileptic patients and elderly depressed patients were prone to repeated falls. These results have the potential to assist in identifying patients at high risk and in designing and implementing strategies to prevent such incidents.     

Evolving Perceptions of Clinical Management in Acute Hospitals in England

Clinical management is at the centre of organizational change in the contemporary British hospital sector, and throws up important questions about the relationship between medicine and management. Through the analysis of business planning as a key organizational process the interpretative frameworks which guide the perceptions of the roles of clinical directors and directorate managers can be understood. This paper presents empirical material from a study of an acute hospital trust going through a full business planning cycle, and traces the evolving interpretations of the nature, content and form of clinical management. It also provides insights into the changing face of the medical profession and its place within the health services domain.     

The rationalization process of online game cheating behaviors

Online deviant behaviors have been prevalent since the Internet became ubiquitous, but research in this field is relatively scarce. The present study investigates a unique deviant behavior in online video gaming communities – game cheating – to further understand the scope of online deviant social behaviors. This paper aims to uncover how game players actively create and negotiate new meanings of game cheating in online games. Six focus groups were conducted to explore the rationalization process for engaging in game cheating among players of different online video games. It was discovered that elements of intentionality, utility, core gaming goals, evolving gaming norm, and social ties to gaming communities serve as important rationalizing process to justify game cheating behaviors. Through the discovery of these mental schemas for online game cheating, the paper discusses how online deviant behaviors are connected with social life online and form a unique online culture.     

Economic Dependency and HIV/AIDS Prevalence in the Developing World: A Comparative,Longitudinal Analysis

The HIV/AIDS pandemic has plagued global society for over three decades. While breakthroughs in antiretroviral treatments (ART) have proven effective in suppressing the virus and HIV/AIDS intervention outreach have widened, epidemic control remains unevenly achieved among countries. At least 95 percent of HIV/AIDS sufferers originate from developing countries. Dependency theory suggests that developing countries' reliance on debt, trade, and foreign investments pose negative effects on their populations' health. Guided by dependency theory's propositions, this cross‐national study assesses whether increasing dependence on trade, debt, and foreign direct investment potentially increases adult HIV prevalence in developing countries from 1989 to 2012. Using a sample of over 80 nations, we perform a two‐way fixed‐effects OLS regression to evaluate the impact of increasing debt, trade, and foreign investment on adult HIV prevalence. Total debt, short‐term debt, external debt, and GDP were found to increase HIV prevalence. The findings for debt support dependency theory's predictions concerning the ramifications of global economic inequality on HIV/AIDS prevalence.     

Variational inference for sparse spectrum Gaussian process regression

We develop a fast variational approximation scheme for Gaussian process (GP) regression, where the spectrum of the covariance function is subjected to a sparse approximation. Our approach enables uncertainty in covariance function hyperparameters to be treated without using Monte Carlo methods and is robust to overfitting. Our article makes three contributions. First, we present a variational Bayes algorithm for fitting sparse spectrum GP regression models that uses nonconjugate variational message passing to derive fast and efficient updates. Second, we propose a novel adaptive neighbourhood technique for obtaining predictive inference that is effective in dealing with nonstationarity. Regression is performed locally at each point to be predicted and the neighbourhood is determined using a measure defined based on lengthscales estimated from an initial fit. Weighting dimensions according to lengthscales, this downweights variables of little relevance, leading to automatic variable selection and improved prediction. Third, we introduce a technique for accelerating convergence in nonconjugate variational message passing by adapting step sizes in the direction of the natural gradient of the lower bound. Our adaptive strategy can be easily implemented and empirical results indicate significant speedups.     

“Y,what do you think about what X just said?” Conversation analysis of stance-eliciting questions in open dialogue network meetings

Open dialogue is a resource-oriented approach to mental health, which aims to help those involved in a crisis situation support each other and engage in dialogue. While language use in open dialogue is generally characterized as being open-ended, nondirective, and nonevaluating on the professionals' side, little is known about the specific conversational features. The aim of this study was to analyze the interactional functions of a stance-eliciting question of the form: “Y, what do you think about what X just said?” We used conversation analysis (CA) to examine this in eight video-recorded Danish open dialogue network meetings. In CA, stance involves expressions of attitude, affect, or judgments about people or events. We describe the interactional functions of these stance-eliciting questions and discuss how clinicians used them as a tool for accomplishing neutrality. The findings can be used to gain a more nuanced understanding of open dialogue psychotherapeutic practices.     

Parent physical punishment and child aggression in a Singapore Chinese preschool sample

We examine how parental physical punishment (caning and slapping) and child aggression are related, and possible moderation by authoritative control and rejection. A sample of 286 Singapore Chinese preschoolers ages 4–6 reported on rejection; their parents reported on control, caning, and slapping; and their teachers rated child aggression. Results show that father caning is related to aggression, regardless of child gender, whereas mother caning is related to child aggression only at low rejection. Mother slapping is related to sons’ aggression, whereas father slapping is related to daughters’ aggression only at low rejection. Control does not moderate any of the punishment‐aggression links. The punishment‐aggression link is thus a complex one, dependent on the dyad, the punitive act, and the parent's behavior.