Variance component models for longitudinal count data with baseline information: epilepsy data revisited

Random effect models have often been used in longitudinal data analysis since they allow for association among repeated measurements due to unobserved heterogeneity. Various approaches have been proposed to extend mixed models for repeated count data to include dependence on baseline counts. Dependence between baseline counts and individual-specific random effects result in a complex form of the (conditional) likelihood. An approximate solution can be achieved ignoring this dependence, but this approach could result in biased parameter estimates and in wrong inferences. We propose a computationally feasible approach to overcome this problem, leaving the random effect distribution unspecified. In this context, we show how the EM algorithm for nonparametric maximum likelihood (NPML) can be extended to deal with dependence of repeated measures on baseline counts.     

Who Wants to be a Social Worker? Using Routine Published Data to Identify Trends in the Numbers of People Applying for and Completing Social Work Programmes in England

Despite widely reported problems with recruitment and retention,there is surprisingly little published research investigatingthe numbers of people wishing to enter social work and theirdemographic characteristics. This article uses routinely collectedpublished higher education data and reports produced by theGeneral Social Care Council (GSCC) and its predecessor the CentralCouncil Education for Education and Training in Social Work(CCETSW) to look at recent trends in the numbers of people wishingto become social workers. It argues that information on theoverall numbers of people applying each year is less helpfulthan understanding more about which groups are under-representedin social work and why. It also draws attention to some positiveindicators, such as the comparative success of social work inattracting groups who may currently be under-represented inhigher education. In addition to suggesting that we need toknow more about the numbers and types of people applying tobe social workers, it concludes that additional work is requiredin establishing employment patterns among the social work workforceas a whole. For example, a high proportion of newly qualifiedsocial workers take up paid employment in social work, but littleis known about what happens to them at later stages of theircareer.     

Piloting external peer review as a model for performance improvement in third‐sector organizations

Peer review is used widely in the United Kingdom as a means of engendering performance improvement in the statutory sector and is being used increasingly to address the needs of specific networks of third‐sector organizations. The project on which this article is based sought to develop a peer review model that could be applied by any third‐sector organization to improve performance. The article reviews existing initiatives and details elements of models that the research sought to replicate. It describes the approach taken to develop the model, including details of three pilots, and summarizes the associated guidance. Finally, it incorporates a commentary on some of the opportunities and challenges presented by the model as a means of securing continued performance improvement by third‐sector organizations.     

A general maximum likelihood analysis of overdispersion in generalized linear models

This paper presents an EM algorithm for maximum likelihood estimation in generalized linear models with overdispersion. The algorithm is initially derived as a form of Gaussian quadrature assuming a normal mixing distribution, but with only slight variation it can be used for a completely unknown mixing distribution, giving a straightforward method for the fully non-parametric ML estimation of this distribution. This is of value because the ML estimates of the GLM parameters may be sensitive to the specification of a parametric form for the mixing distribution. A listing of a GLIM4 algorithm for fitting the overdispersed binomial logit model is given in an appendix.A simple method is given for obtaining correct standard errors for parameter estimates when using the EM algorithm.Several examples are discussed.     

African-American families and child protection

Two hundred and seventy African-American and Caucasian families referred to child protection for alleged maltreatment were compared to assess the degree to which they were differentially referred to and processed by child protection. Results indicated that, although African-Americans were referred to child protection by different sources than Caucasian families, reporter bias was unlikely to account for the differences. Those African-American families referred to child protection were more likely to come from female headed households and presumed to be poorer as a result. However, once reported, they were neither substantiated at a higher rate nor kept open for services at a higher rate compared to Caucasian families. Finally, there was very little evidence of differential caseworker attention to African-Americans compared with Caucasian families. The study concluded that differential referral source and/or differential worker attention were inadequate explanations for the over-representation of African-Americans in the child protection system.     

New information and social trust: asymmetry and perseverance of attributions about hazard managers.

It has been argued that news about negative events has a much stronger effect on decreasing social trust than does news about positive events on increasing it. This asymmetry principle of trust was investigated in two surveys that also investigated the perseverance of trust. The possibility that established trust attributions persevere in the face of new information raises questions about the limits of trust asymmetry. The two studies yielded evidence that both type of news (good versus bad) and initial general trust in the nuclear power industry or the food supply industry affected level of trust. Compared to individuals trusting the industry, those distrusting the industry exhibited less trust following both bad and good news events. Study I also found that judged informativeness and judged positiveness of news events were affected by type of news and general trust of the industry. Individuals low in general trust of the nuclear power industry judged both bad news and good news as less positive than did those high in general trust. Those low in general trust judged bad news as more informative than good news and than did those high in general trust. An important implication of the perseverance of trust is to focus attention on including not only the effects of information about specific events and actions, but also on the judgment processes underlying social trust. The Salient Value Similarity model is suggested as one way of accounting for these psychological processes.     

Maternal Psychosocial Adversity and the Longitudinal Development of Infant Sleep

Research has identified associations between indicators of social disadvantage and the presence of child sleep problems. We examined the longitudinal development of infant sleep in families experiencing high (n = 58) or low (n = 64) levels of psychosocial adversity, and the contributions of neonatal self‐regulatory capacities and maternal settling strategies to this development. Assessments of infant sleep at 4‐, 7‐, and 12‐weeks postpartum indicated no differences in sleeping difficulties between high‐ and low‐adversity groups. However, more infant sleep difficulties were reported in the high‐ versus low‐adversity groups at 12‐ and 18‐month follow‐ups. Neonatal self‐regulatory capacities were not related to the presence or absence of adversity, or to subsequent infant sleep quality. However, there were group differences in maternal settling strategies that did predict subsequent infant sleep difficulties. The pattern of sleep disturbance observed in association with maternal psychosocial adversity at 18‐months was consistent with risk for broader impairments in child functioning.     

Tracing studies and analysis of the effect of loss to follow-up on mortality estimation from patient registry data

Summary. Before patient registries are used for studies of the long-term mortality that is associated with chronic medical conditions, the potential bias resulting from patients who become lost to follow-up must be investigated. A study design, used for a systemic lupus erythematosus patient registry, is described. The design involves tracing patients who are defined as 'lost to follow-up' according to specific criteria. This provides supplementary information on the mortality experience of patients who are lost to (regular) follow-up. Some methods of analysis are described, based on comparing the mortality experience of patients when under regular follow-up with the experience of patients after they are deemed to be lost to follow-up. The effect of loss to follow-up, death reporting and visits to the clinic on estimation procedures is illustrated and recommendations are made for patient registries which are to be used in mortality studies.     

An analysis of models for the dilution and adulteration of fruit juice

We present models based on the multivariate normal distribution to represent the process of dilution and adulteration of citrius juice. The models specify a common dilution parameter for those components of the juice which are affected by dilution but not adulteration.Statistical testing of the hypothesis of no dilution or adulteration presents theoretical difficulties. These difficulties are resolved by model comparisons based on averaged, rather than maximized, likelihoods.