Cycling safety policy and research have mostly focused on cycling infrastructure, cyclists' behavior, and safety equipment in the past decades. However, the role ofbuilt environment characteristics (BECs) in the safety of cyclists has not yet been fully examined. For the Netherlands, this is rather surprising given the significant modal share of bicycles in daily trips, the importance attributed to urban spatial planning, and it being one of the most planned countries in the world. Despite the considerable improvements that have ta1cen place in traffic safety over the decades, the ( actual) number of cyclist deaths between 2011 and 2020 increased by on average 2% per year; the cyclists bad a major portion oftraffic death (followed by passenger cars); also, almost onethird of traffic death happened in built-up a.reas (about 25% of fatalities occurred on 50km/h roads in urban areas) in this period. Considering the aim of construction of on average 75,000 new homes per year until 2025, as weil as promoting bicycle use in as a healthy and sustainable mode of transport in the N etherlands, underst.anding the relationships between the BECs and cycling safety is invaluable for improving the safety of cyclists.

The classical Poisson, geometric and negative binomial regression models for count data belong to the family of generalized linear models and are available at the core of the statistics toolbox in the R system for statistical computing. After reviewing the conceptual and computational features of these methods, a new implementation of zero-inflated and hurdle regression models in the functions zeroinfl() and hurdle() from the package pscl is introduced. It re-uses design and functionality of the basic R functions just as the underlying conceptual tools extend the classical models. Both model classes are able to incorporate over-dispersion and excess zeros - two problems that typically occur in count data sets in economics and the social and political sciences - better than their classical counterparts. Using cross-section data on the demand for medical care, it is illustrated how the classical as well as the zero-augmented models can be fitted, inspected and tested in practice. (author's abstract) / Series: Research Report Series / Department of Statistics and Mathematics

The classical Poisson, geometric and negative binomial regression models for count data belong to the family of generalized linear models and are available at the core of the statistics toolbox in the R system for statistical computing. After reviewing the conceptual and computational features of these methods, a new implementation of hurdle and zero-inflated regression models in the functions hurdle() and zeroinfl() from the package pscl is introduced. It re-uses design and functionality of the basic R functions just as the underlying conceptual tools extend the classical models. Both hurdle and zero-inflated model, are able to incorporate over-dispersion and excess zeros-two problems that typically occur in count data sets in economics and the social sciences-better than their classical counterparts. Using cross-section data on the demand for medical care, it is illustrated how the classical as well as the zero-augmented models can be fitted, inspected and tested in practice. (authors' abstract)

The increasingly serious pedestrian safety issue in the City of Austin aroused the concern. Other than conducting quantitative analysis at aggregate level via collecting and examining the secondary data extracted from the existing datasets, the authors shifted towards the disaggregate level analysis, focusing on twenty-six hotspots of pedestrian collisions via mixed method research. Qualitative data was collected in the field survey to precisely capture the contextual features of collision locations, and was interpreted and coded as explanatory variables for the quantitative analysis. Instead of the frequency of pedestrian collision, crash rate measured by incident count per million pedestrians was the dependent variable to identify the factors truly influencing the pedestrian safety issue, not just the total number of walkers. The stepwise bivariate analysis and negative binomial regression examined the association between pedestrian collision rate and independent variables. Finally, the average block length, speed limit posted, sidewalk condition, and the degree of proximity to major pedestrian attractors were statistically significant factors correlating with the pedestrian collision risk. / text

Why do some member states infringe EU law more than others? Based on the quantitative and qualitative analysis reported here, is not because of administrative capacity limitations, but because of political context, policy changes and deliberate opposition by member governments in order to maintain their independence. States in turn, are motivated by domestic politics to seek to avoid implementing EU law. Additionally, I find that richer countries violate the law more often than poorer countries. Further, member states infringe more than others because of a high number of institutional and coalitional veto players. These results suggest that member states are in the EU because the EU serves their national interest over collective ones. Finally, these results suggest new hypothesis. Member states that have a high level of public discontent with the EU are unlikely to tolerate the political costs of implementing EU legislation.

Thesis advisor: Heather Rowan-Kenyon / According to the National Center for Health Statistics (2007), 18-24 year olds are most likely to report heavy drinking in the past year compared to other adults. Heavy alcohol use is problematic not only in itself, but also because it is associated with a host of other negative outcomes. Research has identified traditional-age college men (age 18-24), who are White, and members of a Greek organization or athletic team as the most likely to consume alcohol in excess (Ham & Hope, 2003; Hingson & White, 2012). White men, members of Greek organizations, and college athletes are also the populations least likely to change their behavior as a result of current alcohol interventions employed by colleges and universities (Fachini, Aliane, Martinez, & Furtado, 2012; LaBrie, Pedersen, Lamb, & Quinlan, 2007; Lundahl, Kunz, Brownell, Tollefson, & Burke, 2010; Mattern & Neighbors, 2004). The primary shortcoming of previous research into this problem, is that it has failed to take an intersectional approach to the phenomenon of college men’s alcohol use. To address this gap, this study surveyed 1,457 college men across five college in the Northeastern United States, using the Conformity to Masculine Norms Inventory (CMNI; Mahalik et al., 2003) the Revised Multigroup Ethnic Identity Measure (MEIM-R; Phinney & Ong, 2007) and the Social Dominance Orientation scale (SDO; Pratto, Sidanius, Stallworth, & Malle, 1994). Alcohol consumption was predicted using zero-inflated negative binomial regressions and zero-inflated Poisson regressions, and alcohol problems were predicted using logistic regressions. The study found that the college men’s drinking was primarily predicted by the masculine norms of risk taking, having power over women, emotional control, and desiring multiple sexual partners. Although the sample size was smaller, for non-White respondents in the study, men’s drinking was also predicted by a focus on heterosexual presentation, and the SDO factor of group based dominance. Alcohol problems were largely predicted by the same masculine norms. / Thesis (PhD) — Boston College, 2016. / Submitted to: Boston College. Lynch School of Education. / Discipline: Educational Leadership and Higher Education.

Uma suposição muito comum na análise de modelos de regressão é a de respostas independentes. No entanto, quando trabalhamos com dados longitudinais ou agrupados essa suposição pode não fazer sentido. Para resolver esse problema existem diversas metodologias, e talvez a mais conhecida, no contexto não Gaussiano, é a metodologia de Equações de Estimação Generalizadas (EEGs), que possui similaridades com os Modelos Lineares Generalizados (MLGs). Essas similaridades envolvem a classificação do modelo em torno de distribuições da família exponencial e da especificação de uma função de variância. A única diferença é que nessa função também é inserida uma matriz trabalho que inclui a parametrização da estrutura de correlação dentro das unidades experimentais. O principal objetivo desta dissertação é estudar como esses modelos se comportam em uma situação específica, de dados de contagem com sobredispersão. Quando trabalhamos com MLGs esse problema é resolvido através do ajuste de um modelo com resposta binomial negativa (BN), e a ideia é a mesma para os modelos envolvendo EEGs. Essa dissertação visa rever as teorias existentes em EEGs no geral e para o caso específico quando a resposta marginal é BN, e além disso mostrar como essa metodologia se aplica na prática, com três exemplos diferentes de dados correlacionados com respostas de contagem. / An assumption that is common in the analysis of regression models is that of independent responses. However, when working with longitudinal or grouped data this assumption may not have sense. To solve this problem there are several methods, but perhaps the best known, in the non Gaussian context, is the one based on Generalized Estimating Equations (GEE), which has similarities with Generalized Linear Models (GLM). Such similarities involve the classification of the model around the exponential family and the specification of a variance function. The only diference is that in this function is also inserted a working correlation matrix concerning the correlations within the experimental units. The main objective of this dissertation is to study how these models behave in a specific situation, which is the one on count data with overdispersion. When we work with GLM this kind of problem is solved by setting a model with a negative binomial response (NB), and the idea is the same for the GEE methodology. This dissertation aims to review in general the GEE methodology and for the specific case when the responses follow marginal negative binomial distributions. In addition, we show how this methodology is applied in practice, with three examples of correlated data with count responses.

Promover a segurança viária no entorno escolar é uma estratégia que contribui para que sejam construídas cidades seguras, saudáveis e sustentáveis. Nesse sentido, este estudo é dedicado a identificar a influência das características da estrutura urbana na frequência e na severidade dos acidentes no entorno de escolas de educação básica de Porto Alegre. A análise da frequência e da severidade de acidentes foi conduzida através da estimação de modelos econométricos: binomial negativo e logit ordenado, respectivamente. Para esses, foram calculados os efeitos marginais, permitindo a observação da magnitude dos impactos das variáveis explicativas sobre as variáveis dependentes. As variáveis dependentes frequência e severidade foram extraídas dos acidentes registrados em Porto Alegre entre 2012 e 2014. Foram incluídas, simultaneamente, variáveis da estrutura urbana, das escolas, socioeconômicas e dos acidentes (para a severidade). A partir do geoprocessamento dos dados existentes, os entornos escolares puderam ser caracterizados para três diferentes áreas circulares de análise (buffer ring) de raios de 100, 150 e 200 metros, permitindo a comparação do uso das diferentes áreas. O conjunto de estimativas indica que áreas menores produzem modelos de melhor desempenho para ambas as técnicas empregadas. No entanto, áreas maiores permitem a análise de maior quantidade de variáveis relativas à estrutura urbana. Essa relação sugere os benefícios da escolha a partir do trade-off entre ajuste do modelo e sua capacidade de propiciar análises de variáveis. Foi identificado que a frequência e a severidade de acidentes podem estar relacionadas a uma única variável explicativa de formas opostas – a partir de sinais contrários. Essa diferença de resultados para frequência e severidade de acidentes indica que há maiores benefícios em analisá-las em conjunto. Identificou-se ainda que existem benefícios para a segurança viária em áreas de estrutura urbana com quarteirões menores e maior quantidade de interseções de quatro vias, em frequência e severidade, respectivamente. Já as áreas mais arborizadas tendem a apresentar acidentes de menor severidade nos casos de usuários de modos ativos. / Fostering road safety nearby schools is a strategy that contributes to build safe, healthy and sustainable cities. The aim of this study is to identify the influence of the built environment characteristics in the frequency and severity of accidents nearby elementary and secondary schools in Porto Alegre. The frequency and severity of the accidents were analyzed using econometric models: negative binomial and ordered logit, respectively. The evaluation of their marginal effects allowed the magnitude of the impact caused by the explanatory variable on the dependent variables to be observed. The measured variables frequency and severity were extracted from accidents registered in Porto Alegre from 2012 and 2014. Built environment, socioeconomic and school variables were also included, as well as accident data (for severity). Data geoprocessing allowed school surroundings to be characterized for three different buffer rings, measuring 100, 150 and 200 meters of radius. Thereby it was possible to compare the inclusion of different areas in the study. The estimations indicates that models based on smaller areas have better performances for both employed techniques, whereas larger areas allow the study of a bigger quantity of urban infrastructure variables. That indicates the benefits of choosing based on a trade-off between model adjustment and capacity to engender the analysis of variables. It was shown that frequency and severity of accidents could be related to a single explanatory variable in opposite ways – based on contrary signs. This difference in the results found for frequency and severity indicates that there are more benefits when analyzing them together. Moreover, there are benefits for road safety in areas where the city blocks are shorter and where there are more four-way intersections, in frequency and severity of accidents, respectively. Also, areas of more important afforestation tend to decrease the severity of accidents involving users of active modes.

This thesis investigates the relationship between conflict, economic growth, state capacity and natural resources in Sub-Saharan Africa. It contributes to the limited research in this area and empirically examines these relationships using different econometric models. The first empirical chapter uses a panel dataset that covers the period 1997 - 2013 to analyse the effects of economic growth on conflict in Nigeria using the negative binomial model. The findings support the direct relationship between economic growth and conflict in Nigeria. Controlling for other factors, the results indicate that increase in growth rate - measured by annual growth rate of GDP per capita - decreases the expected number of conflicts. The study finds no evidence of a relationship between levels of wealth in a state and the incidence of conflicts. The analysis controls for factors such as spill-over effects from other states and year and state effects. Finally, to address potential concerns that economic growth could be a cause of conflict or that other unobserved factors could confound the relationship between economic growth and conflict, the chapter employs instrumental variable (IV) estimation using percentage change in rainfall as an instrument. The results with the IV estimation are similar to the results without IV in terms of both sign and significance, indicating that the negative effect of economic growth on conflicts is not due to reverse causality or omitted variables. For robustness checks, a Panel Autoregressive model (PVAR) is also employed. The second empirical chapter analyses the effect of conflict on state capacity in Sub-Saharan Africa. State capacity is measured in terms of fiscal and legal capacity. It also looks at the effects of internal and external conflicts on state capacity. The chapter adopts the Ordinary least squared (OLS) and the system generalised methods of moments (GMM) estimation methods to analyse the panel data consisting of 49 Sub-Saharan countries over the period 2000 - 2015. The results suggest that conflicts have a negative and significant effect on state capacity. However, when military expenditure is used as a proxy for state capacity it is found that conflict strengthens state capacity. The results are consistent with theoretical argument that internal conflicts polarise societies and make it more difficult for governments to reach a consensus in investing in state capacity, while external conflicts mobilise domestic population against a common enemy thereby helping in state capacity building. Finally, the third empirical chapter examines the effect of natural resources on conflict onset and duration using discrete choice models with a dataset covering the period 1980 -2016. The results on the duration analysis show that natural resources prolong duration of conflicts. However, it is found that not all natural resources prolong duration of conflicts. Oil production does not seem to affect duration, whereas oil reserves and gas production lengthens the duration. The findings from the onset analysis show that both production and reserves of natural resources increase the risk of conflict onset.

Developing sound or reliable statistical models for analyzing vehicle crashes is very important in highway safety studies. A difficulty arises when crash data exhibit overdispersion. Over-dispersion caused by unobserved heterogeneity is a serious problem and has been addressed in a variety ways within the negative binomial (NB) modeling framework. However, the true factors that affect heterogeneity are often unknown to researchers, and failure to accommodate such heterogeneity in the model can undermine the validity of the empirical results. Given the limitations of the NB regression model for addressing over-dispersion of crash data due to heterogeneity, this research examined an alternative model formulation that could be used for capturing heterogeneity through the use of finite mixture regression models. A Finite mixture of Poisson or NB regression models is especially useful when the count data were generated from a heterogeneous population. To evaluate these models, Poisson and NB mixture models were estimated using both simulated and empirical crash datasets, and the results were compared to those from a single NB regression model. For model parameter estimation, a Bayesian approach was adopted, since it provides much richer inference than the maximum likelihood approach. Using simulated datasets, it was shown that the single NB model is biased if the underlying cause of heterogeneity is due to the existence of multiple counting processes. The implications could be poor prediction performance and poor interpretation. Using two empirical datasets, the results demonstrated that a two-component finite mixture of NB regression models (FMNB-2) was quite enough to characterize the uncertainty about the crash occurrence, and it provided more opportunities for interpretation of the dataset which are not available from the standard NB model. Based on the models from the empirical dataset (i.e., FMNB-2 and NB models), their relative performances were also examined in terms of hotspot identification and accident modification factors. Finally, using a simulation study, bias properties of the posterior summary statistics for dispersion parameters in FMNB-2 model were characterized, and the guidelines on the choice of priors and the summary statistics to use were presented for different sample sizes and sample-mean values.