Critical success factors for merger in the UK voluntary sector

This paper reports on research into mergers in the voluntary sector in the UK from 1988 to 1993. It compares empirically derived critical success factors with success factors derived from previous work on mergers in both the voluntary and private sectors. The study concludes that merger can be an extremely effective strategy for voluntary organisations. For it to be so, it has to be undertaken willingly, with a shared vision between the partners of what the merger will achieve. Cultural differences need to be recognised and handled within a trusting context in which mutual respect is backed by an equitable process. Finally; sensitive handling of individuals caught up in the merger process is essential—but something which the voluntary sector ought to be well capable of doing.     

Statistical inference using bootstrap confidence intervals

Bootstrap confidence intervals provide a way of quantifying the uncertainties in the inferences that can be drawn from a sample of data. The idea is to use a simulation, based on the actual data, to estimate the likely extent of sampling error. Michael Wood explains how simple bootstrapping works and explores some of its advantages.     

Families, orphans and children under stress in Zimbabwe

This paper evaluates a programme called FOCUS in Zimbabwe: Families, Orphans, and Children under Stress. The programme was initiated in 1993 and works with seven community groups, providing support to around 6500 children in 2000 households. The programme succeeds through committed community volunteers, who made over 93,000 visits to households in 1999. Communities are mobilised by building on existing practices, using natural entry points, and careful selection of local volunteers and supervisors. There is good community ownership, extremely low volunteer turnover, and reporting systems verify that aims and objectives are met. The programme maintains good ethics, addresses a range of children's rights, and is sustainable through a mixture of donor funds, FACT (sponsoring organisation) support and community support. Issues of concern include better recognition of psycho-social needs, prevention and protection from sexual abuse, reducing stigma and discrimination, increasing child-centred approaches and participation, and reducing gender imbalance among volunteers.     

A Multistate model of fecundability and sterility

This paper develops a multistate hazards model for estimating fecundability and sterility from data on waiting times to conception. Important features of the model include separate sterile and nonsterile states, a distinction between preexisting sterility and sterility that begins after initiation of exposure, and log-normally distributed fecundability among nonsterile couples. Application of the model to data on first birth intervals from Taiwan, Sri Lanka, and the Amish shows that heterogeneity in fecundability is statistically significant at most ages, but that preexisting sterility and new sterility are unimportant before age 40. These results suggest that sterility may not be an important determinant of natural fertility until later reproductive ages.     

A Dimensional CLT for Non-Central Wilks' Lambda in Multivariate Analysis

Abstract.  We consider the non-central distribution of the classical Wilks' lambda statistic for testing the general linear hypothesis in MANOVA. We prove that as the dimension of the observation vector goes to infinity, Wilks' lambda obeys a central limit theorem under simple growth conditions on the non-centrality matrix. In one case we also prove a stronger result: the saddlepoint cumulative distribution function (CDF) approximation for the standardized version of Wilks' lambda converges uniformly on compact sets to the standard normal CDF. These theoretical results go some way towards explaining why saddlepoint approximations to the distribution of Wilks' lambda retain excellent accuracy in high-dimensional cases.     

Thin plate regression splines

Summary. I discuss the production of low rank smoothers for d  ≥ 1 dimensional data, which can be fitted by regression or penalized regression methods. The smoothers are constructed by a simple transformation and truncation of the basis that arises from the solution of the thin plate spline smoothing problem and are optimal in the sense that the truncation is designed to result in the minimum possible perturbation of the thin plate spline smoothing problem given the dimension of the basis used to construct the smoother. By making use of Lanczos iteration the basis change and truncation are computationally efficient. The smoothers allow the use of approximate thin plate spline models with large data sets, avoid the problems that are associated with 'knot placement' that usually complicate modelling with regression splines or penalized regression splines, provide a sensible way of modelling interaction terms in generalized additive models, provide low rank approximations to generalized smoothing spline models, appropriate for use with large data sets, provide a means for incorporating smooth functions of more than one variable into non-linear models and improve the computational efficiency of penalized likelihood models incorporating thin plate splines. Given that the approach produces spline-like models with a sparse basis, it also provides a natural way of incorporating unpenalized spline-like terms in linear and generalized linear models, and these can be treated just like any other model terms from the point of view of model selection, inference and diagnostics.