Mixture models for capture-recapture count data

The contribution investigates the problem of estimating the size of a population, also known as the missing cases problem. Suppose a registration system is targeting to identify all cases having a certain characteristic such as a specific disease (cancer, heart disease, ...), disease related condition (HIV, heroin use, ...) or a specific behavior (driving a car without license). Every case in such a registration system has a certain notification history in that it might have been identified several times (at least once) which can be understood as a particular capture-recapture situation. Typically, cases are left out which have never been listed at any occasion, and it is this frequency one wants to estimate. In this paper modelling is concentrating on the counting distribution, e.g. the distribution of the variable that counts how often a given case has been identified by the registration system. Besides very simple models like the binomial or Poisson distribution, finite (nonparametric) mixtures of these are considered providing rather flexible modelling tools. Estimation is done using maximum likelihood by means of the EM algorithm. A case study on heroin users in Bangkok in the year 2001 is completing the contribution.

Extending Zelterman’s approach for robust estimation of population size to zero-truncated clustered data

Estimation of population size with missing zero-class is an important problem that is encountered in epidemiological assessment studies. Fitting a Poisson model to the observed data by the method of maximum likelihood and estimation of the population size based on this fit is an approach that has been widely used for this purpose. In practice, however, the Poisson assumption is seldom satisfied. Zelterman (1988) has proposed a robust estimator for unclustered data that works well in a wide class of distributions applicable for count data. In the work presented here, we extend this estimator to clustered data. The estimator requires fitting a zero-truncated homogeneous Poisson model by maximum likelihood and thereby using a Horvitz-Thompson estimator of population size. This was found to work well, when the data follow the hypothesized homogeneous Poisson model. However, when the true distribution deviates from the hypothesized model, the population size was found to be underestimated. In the search of a more robust estimator, we focused on three models that use all clusters with exactly one case, those clusters with exactly two cases and those with exactly three cases to estimate the probability of the zero-class and thereby use data collected on all the clusters in the Horvitz-Thompson estimator of population size. Loss in efficiency associated with gain in robustness was examined based on a simulation study. As a trade-off between gain in robustness and loss in efficiency, the model that uses data collected on clusters with at most three cases to estimate the probability of the zero-class was found to be preferred in general. In applications, we recommend obtaining estimates from all three models and making a choice considering the estimates from the three models, robustness and the loss in efficiency. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)

Educational participation post-16: a longitudinal analysis of intentions and outcomes

The issue of levels of participation in post-compulsory education has been emphasised by the current policy initiatives to increase the age to which some form of participation is compulsory. One of the acknowledged weaknesses of research in the field of children's intentions with regard to participation is the lack of longitudinal data. This paper offers a longitudinal analysis using the Youth Survey from the British Household Panel Survey. The results show that most children can express intentions with regard to future participation very early in their secondary school careers and that these intentions are good predictors of actual behaviour five years later. Intentions to stay on are more consistent than intentions to leave and most children who finally leave at 16 have at some point said they want to remain in education post-16. The strongest association with participation levels is attainment at GCSE. However, there are also influences of gender and parental background and these remain, even after attainment is held constant. The results show the value of focusing on intentions for participation at a very early stage of children's school careers and also the importance of current attempts to reform curriculum and assessment for the 14-19 age group.

Longitudinal investigation of the faecal microbiota of healthy full-term infants using fluorescence in situ hybridization and denaturing gradient gel electrophoresis.

From birth onwards, the gastrointestinal (GI) tract of infants progressively acquires a complex range of micro-organisms. It is thought that by 2 years of age the GI microbial population has stabilized. Within the developmental period of the infant GI microbiota, weaning is considered to be most critical, as the infant switches from a milk-based diet (breast and/or formula) to a variety of food components. Longitudinal analysis of the biological succession of the infant GI/faecal microbiota is lacking. In this study, faecal samples were obtained regularly from 14 infants from 1 month to 18 months of age. Seven of the infants (including a set of twins) were exclusively breast-fed and seven were exclusively formula-fed prior to weaning, with 175 and 154 faecal samples, respectively, obtained from each group. Diversity and dynamics of the infant faecal microbiota were analysed by using fluorescence in situ hybridization and denaturing gradient gel electrophoresis. Overall, the data demonstrated large inter- and intra-individual differences in the faecal microbiological profiles during the study period. However, the infant faecal microbiota merged with time towards a climax community within and between feeding groups. Data from the twins showed the highest degree of similarity both quantitatively and qualitatively. Inter-individual variation was evident within the infant faecal microbiota and its development, even within exclusively formula-fed infants receiving the same diet. These data can be of help to future clinical trials (e.g. targeted weaning products) to organize protocols and obtain a more accurate outline of the changes and dynamics of the infant GI microbiota.

The impact of corporate social performance on financial risk and utility: a longitudinal analysis

This study focuses on the wealth-protective effects of socially responsible firm behavior by examining the association between corporate social performance (CSP) and financial risk for an extensive panel data sample of S&P 500 companies between the years 1992 and 2009. In addition, the link between CSP and investor utility is investigated. The main findings are that corporate social responsibility is negatively but weakly related to systematic firm risk and that corporate social irresponsibility is positively and strongly related to financial risk. The fact that both conventional and downside risk measures lead to the same conclusions adds convergent validity to the analysis. However, the risk-return trade-off appears to be such that no clear utility gain or loss can be realized by investing in firms characterized by different levels of social and environmental performance. Overall volatility conditions of the financial markets are shown to play a moderating role in the nature and strength of the CSP-risk relationship.

Crossbred cow adoption and milk market participation in a multivariate count data framework

Cross-bred cow adoption is an important and potent policy variable precipitating subsistence household entry into emerging milk markets. This paper focuses on the problem of designing policies that encourage and sustain milkmarket expansion among a sample of subsistence households in the Ethiopian highlands. In this context it is desirable to measure households’ ‘proximity’ to market in terms of the level of deficiency of essential inputs. This problem is compounded by four factors. One is the existence of cross-bred cow numbers (count data) as an important, endogenous decision by the household; second is the lack of a multivariate generalization of the Poisson regression model; third is the censored nature of the milk sales data (sales from non-participating households are, essentially, censored at zero); and fourth is an important simultaneity that exists between the decision to adopt a cross-bred cow, the decision about how much milk to produce, the decision about how much milk to consume and the decision to market that milk which is produced but not consumed internally by the household. Routine application of Gibbs sampling and data augmentation overcome these problems in a relatively straightforward manner. We model the count data from two sites close to Addis Ababa in a latent, categorical-variable setting with known bin boundaries. The single-equation model is then extended to a multivariate system that accommodates the covariance between crossbred-cow adoption, milk-output, and milk-sales equations. The latent-variable procedure proves tractable in extension to the multivariate setting and provides important information for policy formation in emerging-market settings

Sources, sinks, and transatlantic transport of North African dust aerosol: a multimodel analysis and comparison with remote sensing data

This study evaluates model-simulated dust aerosols over North Africa and the North Atlantic from five global models that participated in the Aerosol Comparison between Observations and Models phase II model experiments. The model results are compared with satellite aerosol optical depth (AOD) data from Moderate Resolution Imaging Spectroradiometer (MODIS), Multiangle Imaging Spectroradiometer (MISR), and Sea-viewing Wide Field-of-view Sensor, dust optical depth (DOD) derived from MODIS and MISR, AOD and coarse-mode AOD (as a proxy of DOD) from ground-based Aerosol Robotic Network Sun photometer measurements, and dust vertical distributions/centroid height from Cloud Aerosol Lidar with Orthogonal Polarization and Atmospheric Infrared Sounder satellite AOD retrievals. We examine the following quantities of AOD and DOD: (1) the magnitudes over land and over ocean in our study domain, (2) the longitudinal gradient from the dust source region over North Africa to the western North Atlantic, (3) seasonal variations at different locations, and (4) the dust vertical profile shape and the AOD centroid height (altitude above or below which half of the AOD is located). The different satellite data show consistent features in most of these aspects; however, the models display large diversity in all of them, with significant differences among the models and between models and observations. By examining dust emission, removal, and mass extinction efficiency in the five models, we also find remarkable differences among the models that all contribute to the discrepancies of model-simulated dust amount and distribution. This study highlights the challenges in simulating the dust physical and optical processes, even in the best known dust environment, and stresses the need for observable quantities to constrain the model processes.

Metabolic acidosis in patients with severe sepsis and septic shock: A longitudinal quantitative study

Objective: To describe the composition of metabolic acidosis in patients with severe sepsis and septic shock at intensive care unit admission and throughout the first 5 days of intensive care unit stay. Design: Prospective, observational study. Setting: Twelve-bed intensive care unit. Patients: Sixty patients with either severe sepsis or septic shock. Interventions: None. Measurements and Main Results: Data were collected until 5 days after intensive care unit admission. We studied the contribution of inorganic ion difference, lactate, albumin, phosphate, and strong ion gap to metabolic acidosis. At admission, standard base excess was -6.69 +/- 4.19 mEq/L in survivors vs. -11.63 +/- 4.87 mEq/L in nonsurvivors (p < .05); inorganic ion difference (mainly resulting from hyperchloremia) was responsible for a decrease in standard base excess by 5.64 +/- 4.96 mEq/L in survivors vs. 8.94 +/- 7.06 mEq/L in nonsurvivors (p < .05); strong ion gap was responsible for a decrease in standard base excess by 4.07 +/- 3.57 mEq/L in survivors vs. 4.92 +/- 5.55 mEq/L in nonsurvivors with a nonsignificant probability value; and lactate was responsible for a decrease in standard base excess to 1.34 +/- 2.07 mEq/L in survivors vs. 1.61 +/- 2.25 mEq/L in nonsurvivors with a nonsignificant probability value. Albumin had an important alkalinizing effect in both groups; phosphate had a minimal acid-base effect. Acidosis in survivors was corrected during the study period as a result of a decrease in lactate and strong ion gap levels, whereas nonsurvivors did not correct their metabolic acidosis. In addition to Acute Physiology and Chronic Health Evaluation 11 score and serum creatinine level, inorganic ion difference acidosis magnitude at intensive care unit admission was independently associated with a worse outcome. Conclusions: Patients with severe sepsis and septic shock exhibit a complex metabolic acidosis at intensive care unit admission, caused predominantly by hyperchloremic acidosis, which was more pronounced in nonsurvivors. Acidosis resolution in survivors was attributable to a decrease in strong ion gap and lactate levels. (Crit Care Med 2009; 37:2733-2739)

Using non-homogeneous Poisson models with multiple change-points to estimate the number of ozone exceedances in Mexico City

In this paper, we consider some non-homogeneous Poisson models to estimate the probability that an air quality standard is exceeded a given number of times in a time interval of interest. We assume that the number of exceedances occurs according to a non-homogeneous Poisson process (NHPP). This Poisson process has rate function lambda(t), t >= 0, which depends on some parameters that must be estimated. We take into account two cases of rate functions: the Weibull and the Goel-Okumoto. We consider models with and without change-points. When the presence of change-points is assumed, we may have the presence of either one, two or three change-points, depending of the data set. The parameters of the rate functions are estimated using a Gibbs sampling algorithm. Results are applied to ozone data provided by the Mexico City monitoring network. In a first instance, we assume that there are no change-points present. Depending on the adjustment of the model, we assume the presence of either one, two or three change-points. Copyright (C) 2009 John Wiley & Sons, Ltd.

Estimating the number of ozone peaks in Mexico City using a non-homogeneous Poisson model

In this paper, we consider the problem of estimating the number of times an air quality standard is exceeded in a given period of time. A non-homogeneous Poisson model is proposed to analyse this issue. The rate at which the Poisson events occur is given by a rate function lambda(t), t >= 0. This rate function also depends on some parameters that need to be estimated. Two forms of lambda(t), t >= 0 are considered. One of them is of the Weibull form and the other is of the exponentiated-Weibull form. The parameters estimation is made using a Bayesian formulation based on the Gibbs sampling algorithm. The assignation of the prior distributions for the parameters is made in two stages. In the first stage, non-informative prior distributions are considered. Using the information provided by the first stage, more informative prior distributions are used in the second one. The theoretical development is applied to data provided by the monitoring network of Mexico City. The rate function that best fit the data varies according to the region of the city and/or threshold that is considered. In some cases the best fit is the Weibull form and in other cases the best option is the exponentiated-Weibull. Copyright (C) 2007 John Wiley & Sons, Ltd.

Destructive weighted Poisson cure rate models

In this paper, we develop a flexible cure rate survival model by assuming the number of competing causes of the event of interest to follow a compound weighted Poisson distribution. This model is more flexible in terms of dispersion than the promotion time cure model. Moreover, it gives an interesting and realistic interpretation of the biological mechanism of the occurrence of event of interest as it includes a destructive process of the initial risk factors in a competitive scenario. In other words, what is recorded is only from the undamaged portion of the original number of risk factors.

Parameter estimation of Poisson generalized linear mixed models based on three different statistical principles : a simulation study

Generalized linear mixed models are flexible tools for modeling non-normal data and are useful for accommodating overdispersion in Poisson regression models with random effects. Their main difficulty resides in the parameter estimation because there is no analytic solution for the maximization of the marginal likelihood. Many methods have been proposed for this purpose and many of them are implemented in software packages. The purpose of this study is to compare the performance of three different statistical principles - marginal likelihood, extended likelihood, Bayesian analysis-via simulation studies. Real data on contact wrestling are used for illustration.

A modest start, but a steady rise in research use : A longitudinal study of nurses during the first five years in professional life

Background: Newly graduated nurses are faced with a challenging work environment that may impede theirability to provide evidence-based practice. However, little is known about the trajectory of registered nurses’ use ofresearch during the first years of professional life. Thus, the aim of the current study was to prospectively examinethe extent of nurses’ use of research during the first five years after undergraduate education and specifically assesschanges over time.Method: Survey data from a prospective cohort of 1,501 Swedish newly graduated nurses within the nationalLANE study (Longitudinal Analyses of Nursing Education and Entry in Worklife) were used to investigate perceiveduse of research over the first five years as a nurse. The dependent variables consisted of three single itemsassessing instrumental, conceptual, and persuasive research use, where the nurses rated their use on a five-pointscale, from ‘never’ (1) to ‘on almost every shift’ (5). These data were collected annually and analyzed bothdescriptively and by longitudinal growth curve analysis.Results: Instrumental use of research was most frequently reported, closely followed by conceptual use, withpersuasive use occurring to a considerably lower extent. The development over time showed a substantial generalupward trend, which was most apparent for conceptual use, increasing from a mean of 2.6 at year one to 3.6 atyear five (unstandardized slope +0.25). However, the descriptive findings indicated that the increase started onlyafter the second year. Instrumental use had a year one mean of 2.8 and a year five mean of 3.5 (unstandardizedslope +0.19), and persuasive use showed a year one mean of 1.7 and a year five mean of 2.0 (unstandardized slope+0.09).Conclusion: There was a clear trend of increasing research use by nurses during their first five years of practice.The level of the initial ratings also indicated the level of research use in subsequent years. However, it took morethan two years of professional development before this increase ‘kicked in.’ These findings support previousresearch claiming that newly graduated nurses go through a ‘transition shock,’ reducing their ability to useresearch findings in clinical work.

Using Spatiotemporal Methods to Fill Gaps In Energy Usage Interval Data

Researchers analyzing spatiotemporal or panel data, which varies both in location and over time, often find that their data has holes or gaps. This thesis explores alternative methods for filling those gaps and also suggests a set of techniques for evaluating those gap-filling methods to determine which works best.

A competitividade internacional de firmas transnacionais na nova geografia econômica das nações em processo de integração político-econômica regional: uma análise longitudinal de empresas globais em macrorregiões supranacionais

Devido ao fenômeno da globalização, alguns aspectos em Economia Internacional e Política têm sido amplamente debatidos entre os estudiosos de Negócios Internacionais, dada a extensão do seu impacto sobre a competitividade operacional e estratégica das empresas multinacionais. Em conjunto com a realidade mais regional da maioria dos acordos preferenciais de comércio e de investimentos – que promovem uma integração regional mais profunda dos mercados, ao contrário do teórico mercado "global" – as abordagens teóricas mais globalizantes em estratégia de negócios internacionais têm sido mais questionadas. Enquanto alguns estudiosos, como Pankaj Ghemawat, (2007), propõem abordagens para a chamada "semi-globalização"; outros, por exemplo, com Alan Rugman e Alain Verbeke (inter alias 2004, 2007), por outro lado, defendem abordagens regionalistas mais restritas em negócios internacionais e estratégia de empresas internacionais. Tais posições sobre o desempenho das empresas transnacionais, no entanto, não foram amplamente testadas, deixando, assim, outras questões relevantes sem soluções. Assim sendo, identificou-se um espaço na literatura quanto à questão de as regiões – em vez de países individualmente considerados – serem ou não relevantes ao desempenho global das empresas multinacionais e em que medida. Nesse sentido, foi utilizada uma metodologia quantitativa longitudinal a fim de avaliar a evidência histórica da repercussão de presença em regiões selecionadas e/ou países sobre o desempenho das empresas transnacionais. Foram coletados dados no Compustat Global (2009) com vistas a uma análise econométrica de dados em painel. Nossos resultados consistem, brevemente, em três aspectos. Em primeiro lugar, quando ambas as variáveis (país e região) são simultaneamente consideradas influentes sobre o desempenho, existe significância estatística. Em segundo lugar, ao contrastar ambas as variáveis (país e região) entre si, em relação ao maior nível de impacto no desempenho, ainda que tenhamos encontrado relevância estatística para os países individualmente considerados, suspeitou-se de algum problema nos dados brutos. Em terceiro lugar, ao assumir uma correlação positiva entre o desempenho da empresa multinacional e do número de regiões geográficas onde essas corporações possuem operações significativas, foi encontrada significância estatística. Nossa conclusão, portanto, consiste no fato de que, dado que a maioria dos países são signatários de pelo menos um acordo de integração regional, as regiões devem ser o foco principal dos negócios internacionais e estratégia corporativa internacional, tanto nos propósitos teóricos (tendo em vista as conclusões desta pesquisa e a literatura sobre o assunto), quanto nos aspectos práticos (em vez de da customização da gestão e da estratégia para cada país individual).