John W. Mcdonald

University of Southampton · Southampton Statistical Sciences Research Institute (S3RI)

Demonstrating the impact and effectiveness of educational interventions, including medium and high-fidelity simulation, has long been fraught with methodological challenges and ambiguities. This is particularly the case when there are several confounding factors and variables operating in situations where control trials are inappropriate, and inves...

  We propose a latent class model for bivariate binary responses from monozygotic and dizygotic twin pairs. We study the comorbidity of two diseases, where each disease and their association are presumed to be genetically related via latent predispositions. Each twin belongs to one of four latent classes, conceptualized as a 2×2 table and parameter...

Postponing the start of childbearing raises the question of fertility postponed versus fertility foregone. One of the limitations of previous studies of 'How late can you wait?' is that any observed decline in the probability of conception with age could be due to a decline in fecundability with age or due to a decline in coital frequency with age...

This study investigates the patterns of recording birth weight data in retrospective surveys and their influence on birth weight estimates in less developed countries. We hypothesise that the method of reporting birth weight in surveys influences the classification of infants in the low birth weight category. Population-level data from Demographic...

Connections between graphical Gaussian models and classical single-factor models are obtained by parameterizing the single-factor model as a graphical Gaussian model. Models are represented by independence graphs, and associations between each manifest variable and the latent factor are measured by factor partial correlations. Power calculations fo...

This paper describes how Semantic Grid technologies can be used to generate enhanced tools for data collection that provide enabling technologies for interdisciplinary work, thereby enhancing the capacity to address substantive social science research. A feasibility study is presented in which semantic annotation (i.e. machine-processable annotatio...

This paper describes how Semantic Grid technologies can be used to generate enhanced tools for data collection that provide enabling technologies for interdisciplinary work, thereby enhancing the capacity to address substantive social science research. A feasibility study is presented in which semantic annotation (i.e. machine-processable annotatio...

IntroductionA modelling frameworkResults of the initial simulationAdding noise to the observationsIntroducing a contiguity parameterNorthern European migrationConclusion References

The association structure between manifest variables arising from the single-factor model is investigated using partial correlations. The additional insights to the practitioner provided by partial correlations for detecting a single-factor model are discussed. The parameter space for the partial correlations is presented, as are the patterns of si...

In this paper we use visualisation techniques to study the human-computer interaction (HCI) of nursing students using an interactive web-based environment for learning clinical skills.

Stress-testing has become an important topic in retail lending since the introduction of the new Basel II guidelines. Here we use a scenario-based forecasting approach developed explicitly for retail lending in order to provide a suitable stress-testing approach. We first decompose the historical vintage performance data into a maturation function...

This paper describes initial work on developing a semantic annotation system for the augmentation of skills-based learning for Healthcare. Scenario driven skills-based learning takes place in an augmented hospital ward simulation involving a patient simulator known as SimMan. The semantic annotation software enables real-time annotations of these s...

Members of the public are increasingly consulted over health care and research priorities. Patient involvement in determining cancer research priorities, however, has remained underdeveloped. This paper presents the findings of the first consultation to be conducted with UK cancer patients concerning research priorities. The study adopted a partici...

The uniformly most powerful unbiased test of reciprocity compares the observed number of mutual relations to its exact conditional distribution. Metropolis–Hastings algorithms have been proposed for generating from this distribution in order to perform Monte Carlo exact inference. Triad census statistics are often used to test for the presence of n...

Members of the public are increasingly consulted over health care and research priorities. Patient involvement in determining cancer research priorities, however, has remained underdeveloped. This paper presents the findings of the first consultation to be conducted with UK cancer patients concerning research priorities. The study adopted a partici...

This paper describes initial work on developing a semantic annotation system for the augmentation of skills-based learning for Healthcare. Scenario driven skills-based learning takes place in an augmented hospital ward simulation involving a patient simulator known as SimMan. The semantic annotation software enables real-time annotations of these s...

Weight loss and anorexia are commonly reported symptoms in people with advanced cancer. Little is known about patient experience of these phenomena, in particular whether they find them of concern. In this study, the prevalence of weight loss and eating-related concern was evaluated in patients with advanced cancer receiving specialist palliative h...

The classical single-factor model is parametrized as a graphical Gaussian model. The relationship between the classical parametrization of the single-factor model and this alternative parametrization is derived. This relationship provides extra insights into the single-factor model, which facilitates power calculations. The overall power of the fir...

Asymptotic multivariate normal approximations to the joint distributions of edge exclusion test statistics for saturated graphical log-linear models, with all variables binary, are derived. Non-signed and signed square-root versions of the likelihood ratio, Wald and score test statistics are considered. Non-central chi-squared approximations are al...

Many proposed methods for analyzing clustered ordinal data focus on the regression model and consider the association structure within a cluster as a nuisance. However, the association structure is often of equal interest—for example, temporal association in longitudinal studies and association between responses to similar questions in a survey. We...

We analyse a 2^4 table reporting the presence or absence of alcohol dependence and depression in both members of 597 pairs of female monozygotic twins. The statistical analysis is based on a latent class model, formulated as a 2x2 table and parametrized by two marginal univariate genetic dispositions and the dependence ratio between these dispositi...

The ‘impact’ of academic research is typically measured by how much it is read, used and cited, and by how much new work it influences. Services that measure impact work well for journal-based disciplines. Book-based disciplines can now benefit from online tools and methods of impact analysis too. These analyses also predict fruitful directions for...

Resumen Propósito El propósito del estudio era investigar si existían diferencias en los niveles de conocimiento entre psiquiatras en formación y especialistas daneses e ingleses. Esto es importante en el contexto del movimiento libre (y creciente) de mano de obra médica a través de las fronteras de los países de la Unión Europea (EU). Métodos Se...

The quasi-independence model is a generalization of the model of independence for a two-dimensional contingency table. It states that a particular subset of cells satisfies the independence structure whereby the cell expected count is a product of a row effect and a column effect. It is often used for square tables (number of rows equal number of c...

This article has no abstract.

The purpose of the study was to investigate if differences in levels of knowledge existed between Danish and English training and specialist psychiatrists. This is important in the context of the free (and growing) movement of the medical workforce across European Union (EU) countries' borders. A complete balanced two-way factorial study design was...

Asymptotic multivariate normal approximations to the joint distributions of edge exclusion test statistics for saturated graphical Gaussian models are derived. Non-signed and signed square-root versions of the likelihood ratio, Wald and score test statistics are considered. Noncentral chi-squared approximations are also considered for the non-signe...

Delaying the start of childbearing raises the issue of fertility postponed versus fertility foregone. One of the limits of previous studies of ``How late can you wait?'' is the difficulty of controlling for sexual activity. Data on the frequency and timing of intercourse within a menstrual cycle are uncommon. We use such data from the Menstrual Cyc...

This paper explores whether analysis of school library borrowing records can reveal any insights into the reading habits of junior age children. Data in the form of computerised library-borrowing records were collected from a cohort of 90 Year 6 (age 10–11 years) children over the full course of the academic year. Two-thirds of the cohort were boys...

We propose models for longitudinal, or otherwise clustered, ordinal data. The association between subunit responses is characterized by dependence ratios (Ekholm, Smith, and McDonald, 1995, Biometrika 82, 847-854), which are extended from the binary to the multicategory case. The joint probabilities of the subunit responses are expressed as explici...

We develop Metropolis-Hastings algorithms for exact conditional inference, including goodness-of-fit tests, confidence intervals and residual analysis, for binomial and multinomial logistic regression models. We present examples where the exact results, obtained by enumeration, are available for comparison. We also present examples where Monte Carl...

This paper examines the impact of rural-urban migration on under-two mortality in India, using data from the 1992/93 Indian National Family Health Survey. Multilevel logistic models are fitted for mortality in three age groups: neonatal, early post-neonatal, and late post-neonatal and toddler. Migration status was not a significant determinant of m...

We propose a mixture model that combines a discrete-time survival model for analyzing the correlated times between recurrent events, e.g. births, with a logistic regression model for the probability of never experiencing the event of interest, i.e., being a long-term survivor. The proposed survival model incorporates both observed and unobserved he...

Models for a multivariate binary response are parameterized by univariate marginal probabilities and dependence ratios of all orders. The w-order dependence ratio is the joint success probability of w binary responses divided by the joint success probability assuming independence. This parameterization supports likelihood-based inference for both r...

A test of reciprocity is often performed by analysts of social network data. This test corresponds to testing whether a parameter in an exponential family model for the adjacency matrix is zero. The uniformly most powerful unbiased test compares the observed number of mutual relations in the social network to its exact conditional distribution. As...

Generalized log-linear models can be used to describe the association structure and/or the marginal distributions of multivariate categorical responses. We simultaneously model the association structure and marginal distributions using association-marginal (AM) models, which are specially formulated generalized log-linear models that combine two mo...

We propose Metropolis-Hastings sampling methods for estimating the exact conditional p-value for tests of goodness of fit of log-linear models for mortality rates and standardized mortality ratios. We focus on two-way tables, where the required conditional distribution is a multivariate noncentral hypergeometric distribution with known noncentralit...

A two-way contingency table in which both variables have the same categories is termed a symmetric table. In many applications, because of the social processes involved, most of the observations lie on the main diagonal and the off-diagonal counts are small. For these tables, the model of independence is implausible and interest is then focussed on...

It is common to have knowledge about the unknown parameters of a demographic model which restricts the values of the parameters. With such knowledge, it is reasonable to require that our parameter estimates satisfy the parameter constraints and one should use a constrained estimation procedure. We briefly review existing approaches to fitting param...

The form of the exact conditional distribution of a sufficient statistic for the interest parameters, given a sufficient statistic for the nuisance parameters, is derived for a generalized linear model with canonical link. General results for log‐linear and logistic models are given. A Gibbs sampling approach for generating from the conditional dis...

The value of using exact rather than asymptotic tests to measure intermarriage in the United Kingdom is examined. "We develop Markov chain Monte Carlo methods for estimating the exact conditional p-value and the exact distribution of the residuals, for quasi-independence and quasi-symmetry. These methods are used to analyse a sparse 10 x 10 symmetr...

The form of the exact conditional distribution of a sufficient statistic for the interest parameters, given a sufficient statistic for the nuisance parameters, is derived for a generalized linear model with canonical link. General results for log-linear and logistic models are given. A Gibbs sampling approach for generating from the conditional dis...

Square contingency tables arise frequently in social research. Typically, many of the off-diagonal cell counts are small because of the social processes involved. This causes concern about the validity of using asymptotic tests and an exact test should be considered. We develop Markov chain Monte Carlo methods for estimating the exact conditional p...

We propose the use of the mean parameter for regression analysis of a multivariate binary response. We model the association using dependence ratios defined in terms of the mean parameter, the components of which are the joint success probabilities of all orders. This permits flexible modelling of higher-order associations, using maximum likelihood...

A Monte Carlo exact conditional test of quasi-independence in two-way incomplete contingency tables is proposed. The null distribution of a random table under quasiindependence is derived. This distribution depends only on the counts in the cells of interest and not on the counts in the remaining cells. This result is used to improve the efficiency...

this paper, we propose improvements to a naive simulate and reject procedure for generating r Theta c tables under quasi-independence for an arbitrary pattern of fixed cells. Although some of the algorithmic improvements are described for generating under QI for the off-diagonal cells of a square table, the ideas are applicable to other patterns of...

this paper, is the hypothesis of QI for the off-diagonal cells of a r Theta r square table, where the sufficient statistics for the nuisance parameters are x i+ ; x +j and x ii , for i; j = 1; : : : ; r.

Study of the Extent of Exposure to Low Flying Military Aircraft Noise in the United Kingdom.

In this paper random-effects logistic models are used to analyze the effects of the preceding birth interval on postneonatal mortality in Brazil, controlling for the correlation of survival outcomes between siblings. The results are compared to those obtained by using ordinary logistic regression. Family effects are found to be highly significant i...

Event history analysis - the study of individual life histories - has developed rapidly over the past few years. This book illustrates the use of the new techniques at the frontier of the subject. The number of surveys undertaken throughout the world to collect detailed information on the timing of events in individual lives (eg fertility surveys,...

The effects of birth spacing on neonatal and post-neonatal mortality in Brazil were found to be very consistent with models based on data from other South American countries. The model for neonatal mortality simplified to three significant variables, whereas the model for post-neonatal mortality included four significant interactions. PIP Research...

In this paper we examine the relative importance of a number of demographic determinants of infant and early child mortality using information from 39 World Fertility Survey countries. We include sex of the child, age of the mother at the time of the birth, birth order, mother's educational level and a number of indicators of spacing of adjacent bi...

In this paper five socio-economic correlates of infant and child mortality are considered using results from World Fertility Survey enquiries in 28 developing countries. We consider differences in mortality for segments of life during the first five years on each of the following variables: mother’s education, mother’s work status, husband’s occupa...

A two-way contingency table in which both variables have the same categories is termed a symmetric table. In many applications, because of the social processes involved, most of the observations lie on the main diagonal and the off-diagonal counts are small. For these tables, the model of independence is implausible and interest is then focussed on...

Thesis--University of Washington. Vita. Includes bibliographical references (leaves [108]-110).

Incluye índice