Design and analysis for three level hierarchical structures with count response

In numerous intervention studies and education field trials, random assignment to treatment occurs in clusters rather than at the level of observation. This departure of random assignment of units may be due to logistics, political feasibility, or ecological validity. Data within the same cluster or grouping are often correlated. Application of traditional regression techniques, which assume independence between observations, to clustered data produce consistent parameter estimates. However such estimators are often inefficient as compared to methods which incorporate the clustered nature of the data into the estimation procedure (Neuhaus 1993).1 Multilevel models, also known as random effects or random components models, can be used to account for the clustering of data by estimating higher level, or group, as well as lower level, or individual variation. Designing a study, in which the unit of observation is nested within higher level groupings, requires the determination of sample sizes at each level. This study investigates the design and analysis of various sampling strategies for a 3-level repeated measures design on the parameter estimates when the outcome variable of interest follows a Poisson distribution. ^ Results study suggest that second order PQL estimation produces the least biased estimates in the 3-level multilevel Poisson model followed by first order PQL and then second and first order MQL. The MQL estimates of both fixed and random parameters are generally satisfactory when the level 2 and level 3 variation is less than 0.10. However, as the higher level error variance increases, the MQL estimates become increasingly biased. If convergence of the estimation algorithm is not obtained by PQL procedure and higher level error variance is large, the estimates may be significantly biased. In this case bias correction techniques such as bootstrapping should be considered as an alternative procedure. For larger sample sizes, those structures with 20 or more units sampled at levels with normally distributed random errors produced more stable estimates with less sampling variance than structures with an increased number of level 1 units. For small sample sizes, sampling fewer units at the level with Poisson variation produces less sampling variation, however this criterion is no longer important when sample sizes are large. ^ 1Neuhaus J (1993). “Estimation efficiency and Tests of Covariate Effects with Clustered Binary Data”. Biometrics , 49, 989–996^

Hepatitis B and C infections in U.S. incarcerated populations: Prevalence and related mortality

Published reports have consistently indicated high prevalence of serologic markers for hepatitis B (HBV) and hepatitis C (HCV) infection in U.S. incarcerated populations. Quantifying the current and projected burden of HBV and HCV infection and hepatitis-related sequelae in correctional healthcare systems with even modest precision remains elusive, however, because the prevalence and sequelae of HBV and HCV in U.S. incarcerated populations are not well-studied. This dissertation contributes to the assessment of the burden of HBV and HCV infections in U.S. incarcerated populations by addressing some of the deficiencies and gaps in previous research. ^ Objectives of the three dissertation studies were: (1) To investigate selected study-level factors as potential sources of heterogeneity in published HBV seroprevalence estimates in U.S. adult incarcerated populations (1975-2005), using meta-regression techniques; (2) To quantify the potential influence of suboptimal sensitivity of screening tests for antibodies to hepatitis C virus (anti-HCV) on previously reported anti-HCV prevalence estimates in U.S. incarcerated populations (1990-2005), by comparing these estimates to error-adjusted anti-HCV prevalence estimates in these populations; (3) To estimate death rates due to HBV, HCV, chronic liver disease (CLD/cirrhosis), and liver cancer from 1984 through 2003 in male prisoners in custody of the Texas Department of Criminal Justice (TDCJ) and to quantify the proportion of CLD/cirrhosis and liver cancer prisoner deaths attributable to HBV and/or HCV. ^ Results were as follows. Although meta-regression analyses were limited by the small body of literature, mean population age and serum collection year appeared to be sources of heterogeneity, respectively, in prevalence estimates of antibodies to HBV antigen (HBsAg+) and any positive HBV marker. Other population characteristics and study methods could not be ruled out as sources of heterogeneity. Anti-HCV prevalence is likely somewhat higher in male and female U.S. incarcerated populations than previously estimated in studies using anti-HCV screening tests alone without the benefit of repeat or additional testing. Death rates due to HBV, HCV, CLD/cirrhosis, and liver cancer from 1984 through 2003 in TDCJ male prisoners exceeded state and national rates. HCV rates appeared to be increasing and disproportionately affecting Hispanics. HCV was implicated in nearly one-third of liver cancer deaths. ^

The role and impact of transitional care utilization in colorectal cancer

BACKGROUND Although one out of every five gastrointestinal cancer patients needs transitional care (home-based skilled care or placement in skilled nursing or rehabilitation facilities) following treatment, few studies have examined outcomes in this population compared to patients who return home without assistance. This study has two primary goals: 1. To evaluate long-term cancer-specific outcomes in colorectal cancer patients utilizing transitional care compared to those that return home without assistance following therapy 2. To compare results using standard regression techniques and propensity scores. ^ METHODS Patients undergoing curative surgery for colorectal adenocarcinoma will be identified using data from a tertiary care Veterans Administration hospital. Survival and recurrence will then be determined from VA records and the Social Security Death Index. ^ The association between transitional care utilization and overall and disease-free survival will be evaluated using Cox proportional hazards regression to adjust for confounding factors. Predictors of transitional care utilization will be assessed using multiple logistic regression to generate a propensity score which will also be used to assess differences in survival based on transitional care use. ^ POTENTIAL SIGNIFICANCE If transitional care utilization is associated with worse survival and recurrence following therapy then it will be important to subsequently assess the mechanism in order to target interventions to improve outcomes. If there is no difference in cancer-specific outcomes, then this project can potentially highlight benefits of supportive therapy following colorectal cancer resection.^

Índices de calidad de sitio de Populus nigra ‘Italica’ en plantaciones lineales en el noroeste de Chubut

En el noroeste del Chubut, los álamos se encuentran establecidos en cortinas de protección en predios de pastoreo o donde se cultivan pasturas y forrajes en secano. El Populus nigra ‘Italica’ es el clon más difundido en plantaciones lineales, ubicadas en diferentes calidades de sitio. No existen antecedentes sobre crecimiento en altura e Índice de Sitio en la zona. Se seleccionaron ocho sitios de muestreo, donde fueron apeados 24 árboles dominantes, a los que se les realizó análisis fustal. Con los pares de datos de edad-altura, resultantes del análisis de fuste de los árboles muestra, se ajustó el modelo de Chapman-Richards mediante técnicas de regresión no lineal. Se construyó una familia de curvas de crecimiento en altura, según la metodología de curva guía. Las curvas de Índice de Sitio, basadas en altura y edad, se construyeron mediante deducción matemática a partir de la función de crecimiento en altura, tomándose como edad de referencia 25 años a la altura del pecho. Se definieron cinco calidades de sitio, con un rango de 4 metros, con Índices de Sitio comprendidos entre 19 y 35 m.

Regularization for sparsity in statistical analysis and machine learning

Pragmatism is the leading motivation of regularization. We can understand regularization as a modification of the maximum-likelihood estimator so that a reasonable answer could be given in an unstable or ill-posed situation. To mention some typical examples, this happens when fitting parametric or non-parametric models with more parameters than data or when estimating large covariance matrices. Regularization is usually used, in addition, to improve the bias-variance tradeoff of an estimation. Then, the definition of regularization is quite general, and, although the introduction of a penalty is probably the most popular type, it is just one out of multiple forms of regularization. In this dissertation, we focus on the applications of regularization for obtaining sparse or parsimonious representations, where only a subset of the inputs is used. A particular form of regularization, L1-regularization, plays a key role for reaching sparsity. Most of the contributions presented here revolve around L1-regularization, although other forms of regularization are explored (also pursuing sparsity in some sense). In addition to present a compact review of L1-regularization and its applications in statistical and machine learning, we devise methodology for regression, supervised classification and structure induction of graphical models. Within the regression paradigm, we focus on kernel smoothing learning, proposing techniques for kernel design that are suitable for high dimensional settings and sparse regression functions. We also present an application of regularized regression techniques for modeling the response of biological neurons. Supervised classification advances deal, on the one hand, with the application of regularization for obtaining a na¨ıve Bayes classifier and, on the other hand, with a novel algorithm for brain-computer interface design that uses group regularization in an efficient manner. Finally, we present a heuristic for inducing structures of Gaussian Bayesian networks using L1-regularization as a filter. El pragmatismo es la principal motivación de la regularización. Podemos entender la regularización como una modificación del estimador de máxima verosimilitud, de tal manera que se pueda dar una respuesta cuando la configuración del problema es inestable. A modo de ejemplo, podemos mencionar el ajuste de modelos paramétricos o no paramétricos cuando hay más parámetros que casos en el conjunto de datos, o la estimación de grandes matrices de covarianzas. Se suele recurrir a la regularización, además, para mejorar el compromiso sesgo-varianza en una estimación. Por tanto, la definición de regularización es muy general y, aunque la introducción de una función de penalización es probablemente el método más popular, éste es sólo uno de entre varias posibilidades. En esta tesis se ha trabajado en aplicaciones de regularización para obtener representaciones dispersas, donde sólo se usa un subconjunto de las entradas. En particular, la regularización L1 juega un papel clave en la búsqueda de dicha dispersión. La mayor parte de las contribuciones presentadas en la tesis giran alrededor de la regularización L1, aunque también se exploran otras formas de regularización (que igualmente persiguen un modelo disperso). Además de presentar una revisión de la regularización L1 y sus aplicaciones en estadística y aprendizaje de máquina, se ha desarrollado metodología para regresión, clasificación supervisada y aprendizaje de estructura en modelos gráficos. Dentro de la regresión, se ha trabajado principalmente en métodos de regresión local, proponiendo técnicas de diseño del kernel que sean adecuadas a configuraciones de alta dimensionalidad y funciones de regresión dispersas. También se presenta una aplicación de las técnicas de regresión regularizada para modelar la respuesta de neuronas reales. Los avances en clasificación supervisada tratan, por una parte, con el uso de regularización para obtener un clasificador naive Bayes y, por otra parte, con el desarrollo de un algoritmo que usa regularización por grupos de una manera eficiente y que se ha aplicado al diseño de interfaces cerebromáquina. Finalmente, se presenta una heurística para inducir la estructura de redes Bayesianas Gaussianas usando regularización L1 a modo de filtro.

Location of bioelectricity plants in the Madrid Community based on triticale crop: a multicriteria methodology

This paper presents a work whose objective is, first, to quantify the potential of the triticale biomass existing in each of the agricultural regions in the Madrid Community through a crop simulation model based on regression techniques and multiple correlation. Second, a methodology for defining which area has the best conditions for the installation of electricity plants from biomass has been described and applied. The study used a methodology based on compromise programming in a discrete multicriteria decision method (MDM) context. To make a ranking, the following criteria were taken into account: biomass potential, electric power infrastructure, road networks, protected spaces, and urban nuclei surfaces. The results indicate that, in the case of the Madrid Community, the Campiña region is the most suitable for setting up plants powered by biomass. A minimum of 17,339.9 tons of triticale will be needed to satisfy the requirements of a 2.2 MW power plant. The minimum range of action for obtaining the biomass necessary in Campiña region would be 6.6 km around the municipality of Algete, based on Geographic Information Systems. The total biomass which could be made available in considering this range in this region would be 18,430.68 t.

Family firms and the great crisis: evidence from the italian case

Few studies have analyzed how family firms have acted during the global great crisis in comparison to their nonfamily counterparts. This paper tries to fill this gap on the basis of the Italian experience using a sample of almost 4,500 for 2007 and 2010. We study whether family control affects labour productivity, labour costs and competitiveness and if the adoption of performance related pay (PRP) reveals an efficacious strategy to mitigate the effects of the crisis and reduce the gap in competitiveness with respect to nonfamily firms. We use quantile regression techniques to test the heterogeneous role of PRP and pay attention for its possible endogeneity. We have observed that after the outburst of the crisis, the distance in terms of competitiveness of family firms with respect to their nonfamily counterparts has been amplified. We also find that family firms may take advantage from the adoption of incentive schemes, such as PRP, to encourage commitment and motivation from their employees more than nonfamily firms. The positive role of PRP on labour productivity, coupled with a moderate influence of these schemes on wage premiums, enable them to regain competitiveness also under hostile pressures, as those featuring the strong global crisis.

Green turtle somatic growth dynamics in a spatially disjunct Great Barrier Reef metapopulation

The growth dynamics of green sea turtles resident in four separate foraging grounds of the southern Great Barrier Reef genetic stock were assessed using a nonparametric regression modeling approach. Juveniles recruit to these grounds at the same size, but grow at foraging-ground-dependent rates that result in significant differences in expected size- or age-at-maturity. Mean age-at-maturity was estimated to vary from 25-50 years depending on the ground. This stock comprises mainly the same mtDNA haplotype, so geographic variability might be due to local environmental conditions rather than genetic factors, although the variability was not a function of latitudinal variation in environmental conditions or whether the food stock was seagrass or algae. Temporal variability in growth rates was evident in response to local environmental stochasticity, so geographic variability might be due to local food stock dynamics. Despite such variability, the expected size-specific growth rate function at all grounds displayed a similar nonmonotonic growth pattern with a juvenile growth spurt at 60-70 cm curved carapace length, (CCL) or 15-20 years of age. Sex-specific growth differences were also evident with females tending to grow faster than similar-sized males after the Juvenile growth spurt. It is clear that slow sex-specific growth displaying both spatial and temporal variability and a juvenile growth spurt are distinct growth behaviors of green turtles from this stock.

Does money matter in inflation forecasting

This paper provides the most fully comprehensive evidence to date on whether or not monetary aggregates are valuable for forecasting US inflation in the early to mid 2000s. We explore a wide range of different definitions of money, including different methods of aggregation and different collections of included monetary assets. In our forecasting experiment we use two non-linear techniques, namely, recurrent neural networks and kernel recursive least squares regression - techniques that are new to macroeconomics. Recurrent neural networks operate with potentially unbounded input memory, while the kernel regression technique is a finite memory predictor. The two methodologies compete to find the best fitting US inflation forecasting models and are then compared to forecasts from a naive random walk model. The best models were non-linear autoregressive models based on kernel methods. Our findings do not provide much support for the usefulness of monetary aggregates in forecasting inflation.

Business and social entrepreneurs in the UK:gender, context and commitment

Purpose – The purpose of this paper is to investigate what sort of people become social entrepreneurs, and in what way they differ from business entrepreneurs. More importantly, to investigate in what socio-economic context entrepreneurial individuals are more likely to become social than business entrepreneurs. These questions are important for policy because there has been a shift from direct to indirect delivery of many public services in the UK, requiring a professional approach to social enterprise. Design/methodology/approach – Evidence is presented from the Global Entrepreneurship Monitor (GEM) UK survey based upon a representative sample of around 21,000 adults aged between 16 and 64 years interviewed in 2009. The authors use logistic multivariate regression techniques to identify differences between business and social entrepreneurs in demographic characteristics, effort, aspiration, use of resources, industry choice, deprivation, and organisational structure. Findings – The results show that the odds of an early-stage entrepreneur being a social rather than a business entrepreneur are reduced if they are from an ethnic minority, if they work ten hours or more per week on the venture, and if they have a family business background; while they are increased if they have higher levels of education and if they are a settled in-migrant to their area. While women social entrepreneurs are more likely than business entrepreneurs to be women, this is due to gender-based differences in time commitment to the venture. In addition, the more deprived the community they live in, the more likely women entrepreneurs are to be social than business entrepreneurs. However, this does not hold in the most deprived areas where we argue civic society is weakest and therefore not conducive to support any form of entrepreneurial endeavour based on community engagement. Originality/value – The paper's findings suggest that women may be motivated to become social entrepreneurs by a desire to improve the socio-economic environment of the community in which they live and see social enterprise creation as an appropriate vehicle with which to address local problems.

On the comparative performance of socially responsible and Islamic mutual funds

This is the first study to provide comprehensive analyses of the relative performance of both socially responsible investment (SRI) and Islamic mutual funds. The analysis proceeds in two stages. In the first, the performance of the two categories of funds is measured using partial frontier methods. In the second stage, we use quantile regression techniques.By combining two variants of the Free Disposal Hull (FDH) methods (order-m and order-?) in the first stage of analysis and quantile regression in the second stage, we provide detailed analyses of the impact of different covariates across methods and across different quantiles. In spite of the differences in the screening criteria and portfolio management of both types of funds, variation in the performance is only found for some of the quantiles of the conditional distribution of mutual fund performance. We established that for the most inefficient funds the superior performance of SRI funds is significant. In contrast, for the best mutual funds this evidence vanished and even Islamic funds perform better than SRI.These results show the benefits of performing the analysis using quantile regression.

Estimation and inference under economic restrictions

Estimation of economic relationships often requires imposition of constraints such as positivity or monotonicity on each observation. Methods to impose such constraints, however, vary depending upon the estimation technique employed. We describe a general methodology to impose (observation-specific) constraints for the class of linear regression estimators using a method known as constraint weighted bootstrapping. While this method has received attention in the nonparametric regression literature, we show how it can be applied for both parametric and nonparametric estimators. A benefit of this method is that imposing numerous constraints simultaneously can be performed seamlessly. We apply this method to Norwegian dairy farm data to estimate both unconstrained and constrained parametric and nonparametric models.

On the comparative performance of socially responsible and Islamic mutual funds

This is the first study to provide comprehensive analyses of the relative performance of both socially responsible investment (SRI) and Islamic mutual funds. The analysis proceeds in two stages. In the first, the performance of the two categories of funds is measured using partial frontier methods. In the second stage, we use quantile regression techniques. By combining two variants of the Free Disposal Hull (FDH) methods (order- m and order- α) in the first stage of analysis and quantile regression in the second stage, we provide detailed analyses of the impact of different covariates across methods and across different quantiles. In spite of the differences in the screening criteria and portfolio management of both types of funds, variation in the performance is only found for some of the quantiles of the conditional distribution of mutual fund performance. We established that for the most inefficient funds the superior performance of SRI funds is significant. In contrast, for the best mutual funds this evidence vanished and even Islamic funds perform better than SRI. These results show the benefits of performing the analysis using quantile regression. © 2013 Elsevier B.V.

The relationship of immigrant status to perceptions of reading and reading literacy among young Black students: A test of the Cultural -Ecological Theory of School Performance

This study tests Ogbu and Simons' Cultural-Ecological Theory of School Performance using data from the Progress in International Reading Literacy Study of 2001 (PIRLS), a large-scale international survey and reading assessment involving fourth grade students from 35 countries, including the United States. This theory argues that Black immigrant students outperform their non-immigrant counterparts, academically, and that achievement differences are attributed to stronger educational commitment in Black immigrant families. Four hypotheses are formulated to test this theory: Black immigrant students have (a) more receptive attitudes toward reading; (b) a more positive reading self-concept; and (c) a higher level of reading literacy. Furthermore, (d) the relationship of immigrant status to reading perceptions and literacy persists after including selected predictors. These hypotheses are tested separately for girls and boys, while also examining immigrant students' generational status (i.e., foreign-born or second-generation). ^ PIRLS data from a subset of Black students (N=525) in the larger U.S. sample of 3,763 are analyzed to test the hypotheses, using analysis of variance, correlation and multiple regression techniques. Findings reveal that hypotheses a and b are not confirmed (contradicting the Cultural-Ecological Theory) and c and d are partially supported (lending partial support to the theory). Specifically, immigrant and non-immigrant students did not differ in attitudes toward reading or reading self-concept; second-generation immigrant boys outperformed both non-immigrant and foreign-born immigrant boys in reading literacy, but no differences were found among girls; and, while being second-generation immigrant had a relatively stronger relationship to reading literacy for boys, among girls, selected socio-cultural predictors, number of books in the home and length of U.S. residence, had relatively stronger relationship to reading self-concept than did immigrant status. This study, therefore, indicates that future research employing the Cultural-Ecological Theory should: (a) take gender and generational status into account (b) identify additional socio-cultural predictors of Black children's academic perceptions and performance; and (c) continue to build on this body of evidence-based knowledge to better inform educational policy and school personnel in addressing needs of all children. ^

Regional diversification and performance: Conceptualization and competing strategic objectives

This dissertation comprised of three essays provides justification for the need to pursue research on multinationality and performance with a more fine-grained approach. Essay one is a conceptual response to an article written by Jean-Francois Hennart in 2011 which questions the need and approach toward future research in this domain. I argue that internalization theory does not render multinationality and performance research meaningless and identify key areas where methodological enhancements can be made to strengthen our research findings with regard to Hennart's call for more content validity. Essay two responds to the need for more-fine grained research on the consequences of multinationality by introducing non-traditional measures of performance such as social and environmental performance and adopting a more theoretically relevant construct of regionalization to capture international diversification levels of the firm. Using data from the world's largest 600 firms (based on sales) derived from Bloomberg and the Directory of Corporate Affiliates; I employ general estimating equation analysis to account for the auto-correlated nature of the panel data alongside multivariate regression techniques. Results indicate that regionalization has a positive relationship with economic performance while it has a negative relationship with environmental and social performance outcomes, often referred to as the "Triple Bottom-Line" performance. Essay three builds upon the work in the previous essays by linking the aforementioned performance variables and sample to corporate reputation which has been shown to be a beneficial strategic asset. Using Structural Equation Modeling I explore economic, environmental and social signals as mediators on relationship between regionalization and firm reputation. Results indicate that these variables partially mediate a positive relationship between regionalization and firm reputation. While regionalization positively affects the reputation building signal of economic performance, it aids in reputation building by reducing environmental and social disclosure effects which interestingly impact reputation negatively. In conclusion, the dissertation submits opportunities for future research and contributes to research by demonstrating that regionalization affects performance, but the effect varies in accordance with the performance criterion and context. In some cases, regional diversification may produce competing or conflicting outcomes among the potential strategic objectives of the firm.