This study examines the impact of the COVID-19 pandemic on residents' commuting patterns in the United States. Using the MNL (Multinomial Logit) and binary logit models, we analyze the factors influencing the choice of commuting modes before and during the pandemic. Our findings indicate that various personal, travel-related, and COVID-19-related factors significantly affect commuting choices. For commuting methods other than driving, factors such as gender, age, possession of a driver's license, bicycle ownership, car ownership, family size, working days per week, COVID-19 testing, and mask restrictions play a significant role. The decision to walk to work is influenced by gender, vehicle ownership, travel time, travel distance, working days per week, family income, COVID-19-related relocation, and level of COVID-19 anxiety. Public transportation choices are influenced by factors such as age, race, possession of a driver's license, car ownership, travel time, travel distance, COVID-19-related migration, and COVID-19 testing of cohabitants. Furthermore, the binary logit model reveals that personal factors (e.g., gender, driver's license) and COVID-19-related factors (e.g., mask restrictions, level of concern about the coronavirus) significantly impact the consistency of travel modes before and during the pandemic. This study contributes to our understanding of the changes in commuting patterns during the COVID-19 pandemic and provides insights into the factors that shape residents' commuting choices. The findings can inform transportation planning and policy-making to promote sustainable and resilient transportation systems in the face of future disruptions.

Made available in DSpace on 2016-06-02T20:06:06Z (GMT). No. of bitstreams: 1 4329.pdf: 1652803 bytes, checksum: aad7f31afeb38bcd450c35fd9ed227be (MD5) Previous issue date: 2008-09-02 / Universidade Federal de Sao Carlos / The transactions volume increase brought the fraud increase, which result in a annual loss of billions of reais to all .nancial institutions in the world. Therefore, it.s very important the development of detection methods and fraud prevention. The di¢ cult in modeling this kind of data due the fact the data sets are extremely unbalanced. In this work, a bounded logit model will be proposed for fraud detection. It will also be discussed state- dependent sampling and compared with logit and bounded logit model performances. Two applications, one with a simulated data set and another with a real data set, will be presented. The Bayesian approach to these models will also be discussed. The data set analyses will be implemented in SAS and Winbugs software. / O aumento no volume de transações com cartões de crédito trouxe como consequência o aumento do número de fraudes, o que acarreta em uma perda de bilhões de reais anu- almente à todas instituições .nanceiras do mundo. Com isso é muito importante que metodologias de detecção e prevenção à fraude sejam desenvolvidas. A grande di.cul- dade na modelagem deste tipo de dados é que estes são extremamentes desbalanceados. Neste presente trabalho, será proposto o modelo logito limitado na detecção de fraude. Também será discutido as amostras do tipo state-dependent e comparado os desempenhos dos modelos logito e logito limitado. Duas aplicações, uma com um conjunto de dados simulados e outra com um conjunto de dados reais, serão apresentadas. A abordagem bayesiana para estes modelos também será discutida. As análises dos conjuntos de dados serão realizadas nos softwares SAS e Winbugs.

The equine influenza virus is a significant cause of respiratory disease in horses. Even though horses generally recover from this virus, sometimes horses with equine influenza develop secondary bacterial infections which can cause severe pneumonia, thereby increasing recovery times. Owners and managers are faced with the decision of whether to delay preventative treatment in hopes of the horse avoids contracting a secondary bacterial infection (“wait and see”) or aggressively treat the horse with an antibiotic in hopes of avoiding a serious infection (“treat now”). From a decision making standpoint, the economic considerations include explicit treatment costs as well as nonmonetary costs the owner or manager bear when caring for an ill horse. This study investigating horse owner/manager preferences for treatment alternatives is approached in two parts. The first part of the study collects data from field practitioners to estimate the cost of treatment strategies under different scenarios. The second part consists of a questionnaire presented to horse owners and managers and includes four choices between alternative treatment strategies. Analyzing the data using a conjoint analysis approach, respondents’ willingness to pay for different elements of a treatment strategy are estimated. Based on treatment strategies and demographic interactions, a respondent was willing to pay to cover the cost of a horse who became ill with the equine influenza, but individual price sensitivities suggested horse owners and managers are willing to “treat now” versus “wait and see” in order to not see their horse feel poorly and miss training time.

Bus is the main mode of urban transport in most cities in developing countries. Despite a high mode share, bus service quality is often poor and para-transit services are regarded as a problem in urban transport systems rather than a solution. Using Dhaka as a case study, this thesis investigates bus service quality through identification and valuation of thirteen important attributes using discrete choice models. The attributes examined are travel time, travel cost, waiting time, headway, priority seats for women, crowding inside the bus, boarding and alighting, picking up and dropping off passengers, bus stop facilities, driving quality, driver and crew behaviour, cleanliness inside the bus, and air conditioning. Five focus groups were conducted to identify key qualitative bus attributes and their levels in order to design choice experiments for valuation. A survey of 431 respondents in Dhaka was then undertaken. Two choice experiments were designed and implemented within the survey, each with seven attributes (set A and set B) with travel cost as the common attribute. Multinomial Logit (MNL) models and Mixed Logit (MXL) models were developed using the Dhaka choice data. Twelve of the thirteen attributes were statistically significant at the 99% level. The values of in-vehicle time (IVT), waiting time and headway were BDT 34.80, 47.40 and 64.20 per hour respectively for low income groups in the segmented model. Waiting time has a premium valuation, 1.36 times higher than IVT, which endorses existing evidence. The highest valuation is for the dummy variable seating all the way which is BDT 42.20 for high income females. The next largest was bus stops properly, picks and drops passengers nicely , followed by wide door and mild steps for boarding and alighting , smooth and safe journey , bus stop with shed, but no seating arrangements , and air conditioning . The lowest value was BDT 4.61 for deck and seats are clean and tidy , for the low income group. The WTP for the qualitative attributes is high, but given the poor level of the existing service and low fare levels this seems reasonable. Income has a significant impact on travel cost, as well as gender on priority seats for women and crowding inside the bus. However, household car ownership does not have a significant impact on any of the bus attributes examined. The high income group has 75% higher WTP for A set attributes and 79% higher WTP for B set attributes than low income group. Females have 76% higher WTP for standing comfortably all the way , but 38% higher WTP for seating all the way compared to the male. However, females have a WTP of BDT 0.44 for per percent of priority seats for women in contrast with males who have a WTP of BDT -0.11. There is significant taste heterogeneity for both quantitative and qualitative attributes. The qualitative attributes for picking up and dropping off passengers, boarding and alighting facilities and driving facilities have higher valuation and this attributes came from the existing within the market competition structure in a highly fragmented bus market. Therefore, it is recommended to introduce competition for the market and incentives for bus industry consolidation.

Nanotechnology is one of the key innovative technologies in the present century. The food industry has applied this technology in each of its sectors. Nanotechnology has tremendous potential in food and agriculture, including advancing agricultural cultivation and food production, enhancing food nutrition and flavor, and improving food packaging and preservation. However, the novel properties of nanoscale materials that allow beneficial applications are also accompanied with uncertainties, even unknown risks. A number of studies have examined public understanding as well as acceptance of nanotechnology via surveys in both the US and Europe. However, most of these studies concentrated on public attitudes in general. Few works focused on specific products, let alone food or food related products. This project will contribute to the literature by calculating monetary valuations (i.e., willingness-to-pay) for canola oil where new techniques are utilized. Using choice experiment survey data, consumers’ valuations for nano attributes were estimated with choice models. As implied, consumers were willing to pay $0.95 less for a typical bottle (48 fl. oz.) of canola oil if it was produced from nanoscale-modified seed; $0.51 less if the final products were packed with nanotechnology-enhanced packaging technique; and no significant difference was found for oil that was designed with health enhancing nano-engineered oil drops, which would require interaction with the human digestive system. Additionally, the results revealed unobserved heterogeneities among respondents in their willingness-to-pay for canola oil attributes. Aligned with descriptive results, 46.7% of the respondents reported that they were optimistic about new technology applied to food products. While a significant portion of the respondents (42.8%) indicated that they might gain benefits at the same level as risks, there were a slightly larger proportion of the respondents who feared they might be exposed to more risks than benefits through nanofoods. Further analysis included respondents’ attitudes and opinions as well as their demographic and socioeconomic characteristics toward the goal of understanding the underlying behavior difference. Findings from this study will help bridge the gap between scientific innovation and public policy and social-economic concerns. Implications for government policy that can be efficiently used to monitor and regulate these technologies were also investigated.

本研究以羅吉斯迴歸模型探討影響上市櫃公司選擇公開募集或私募之因素。 實證結果發現採私募或現金增資的決策，和下列因素有關：(1)資訊不對稱程度愈高之公司愈傾向以私募發行新股，且上櫃公司較上市公司更傾向以私募發行新股；(2) 資訊透明度越低的公司，資金成本越高，籌資也較困難，會傾向以私募方式發行新股；(3) 投資機會愈多的公司，因為控制權考量，會選擇公開募集資金；(4) 當景氣愈好，公司愈傾向以公開募集發行新股，且上市公司較上櫃公司更易傾向以現金增資方式發行新股。 / This study applies logit model to examine the determinants of choice of private placements/seasoned equity offerings. The results indicate that firms being severe in information asymmetry tend to choose private placements. Contrast to listed firms, OTC firms prefer private placements to seasoned equity offerings. Firms with lower information transparency pay higher costs in seasoned equity offerings and therefore tend to choose private placements. Firms being abundant in investment opportunities prefer to choose seasoned equity offerings because of control considerations. In bull markets, more firms choose seasoned equity offerings other than private placements while listed firms have stronger tendency to choose seasoned equity offerings than OTC firms.

為探討台灣有線電視收視戶對「分級付費制度」之接受意願，本文參考相關文獻，建立台灣有線電視收視戶對「分級付費制度」接受意願的邏輯式模型（logit model），並採用中央通訊社於2000年8月針對台灣有線電視收視戶進行電話訪查之調查資料進行實證分析，先以圖表進行交叉分析，再進行邏輯式迴歸分析，探討消費者人口統計變數（如性別、教育程度等）、收視行為資料（如每日收看有線電視時數、最常收看頻道數等）對「分級付費制度」接受意願之影響程度。依實證結果，男性、教育程度較高、年齡較大、每日收看有線電視時數較低、收看頻道數目較少及目前繳交有線電視費用較高之收視戶，較有意願接受「分級付費制度」。 / To analyze Taiwan cable TV viewers’ willingness to accept the tiering system, this study first attempts to build a theoretical model of cable TV viewers’ willingness to accept the tiering system based on prior related researches. Secondly, this study applies the logit model to estimate cable TV viewers’ willingness to accept the tiering system based on Central News Agency (CNA) 2000 survey data of cable TV viewers. The empirical results show that male viewers are more willing to accept the tiering system than female. Besides, age, education and current monthly cable TV fee have significantly positive impact on cable TV viewers’ willingness to accept the tiering system while daily watching hours and the number of channels viewed have significantly negative impacts.

A common problem in regression analysis is that some covariates are measured with errors. In this dissertation we present simulation-based approach to estimation in two popular regression models with a categorical response variable and classical measurement errors in covariates. The first model is the regression model with a binary response variable. The second one is the proportional odds regression with an ordinal response variable. In both regression models we consider method of moments estimators for therein unknown parameters that are defined via minimizing respective objective functions. The later functions involve multiple integrals and make obtaining of such estimators unfeasible. To overcome this computational difficulty, we propose Simulation-Based Estimators (SBE). This method does not require parametric assumptions for the distributions of the unobserved covariates and error components. We prove consistency and asymptotic normality of the proposed SBE's under some regularity conditions. We also examine the performance of the SBE's in finite-sample situations through simulation studies and two real data sets: the data set from the AIDS Clinical Trial Group (ACTG175) study for our logistic and probit regression models and one from the Adult Literacy and Life Skills (ALL) Survey for our regression model with the ordinal response variable and mismeasured covariates. / October 2016

Ces dernières années, le développement continu des marques de distributeurs (MDD) a abouti à un marché composés de trois grands segments de consommateurs : (i) ceux qui sont fidèles aux marques nationales, (ii) aux marques de distributeurs et (iii) ceux qui combinent les deux. Pourquoi ce dernier groupe de consommateurs migrent-t-il d’une marque nationale vers une marque de distributeurs et inversement ? A notre connaissance, aucune étude n’a été menée sur cet aspect de la concurrence entre marque nationale et MDD. Les travaux de recherche en marketing ont surtout étudié les changements entre marques nationales ou le choix des MDD. Or, le comportement migratoire entre les deux types de marques peut représenter jusqu’à 20% des comportements d’achat (source : Panel MarketingScan). L’objet de cette thèse est donc de proposer un cadre théorique permettant de comprendre la migration entre les deux types de marques. Nous analysons l’influence des variables relatives à la marque (ou des références) et à la catégorie de produits tout en tenant compte des différences individuelles (observées et non observées) entre les ménages et les enseignes. Nous utilisons comme cadre empirique le panel Angevin de la Société MarketingScan. Nous élaborons pour cela un modèle de choix avec coefficients aléatoires et facteurs latents (Latent Factor Random coefficients Multinomial Logit Model) permettant de contrôler l’hétérogénéité entre les ménages. Les résultats montrent que le type de MDD (marque enseigne ou marque propre) a une influence à la fois sur la migration vers et le rachat des MDD. Le prix reste toujours important dans la concurrence entre les deux types de marques. Une plus grande disponibilité des références au niveau de la marque de distributeurs permet d’attirer plus de consommateurs. Toutefois, une forte présence de MDD réduit la satisfaction des consommateurs à l’égard de l’assortiment et les pousser à migrer vers les marques nationales. / For the last years, the consistent development of store brands or private labels has resulted in a market composed of three segments: customers who are national brand loyal, (ii) store brand loyal and (iii) those who combine store and national brands. So the question becomes: why does the latter group of consumers migrate from a national brand to a store brand and vice versa? To our knowledge, no study has been conducted about this aspect of national brand and private label competition. The research in marketing has mainly studied brand switching between national brands and the choice of store brand. However, migration between the two types of brand might account for up to 20% of the purchase behaviours (source: MarketingScan panel). The purpose of this dissertation is to suggest a theoretical framework to understand the migration between the two types of brands. We analyzed the influence of variables at the brand and SKU, and product category levels, while accounting for the (observed and unobserved) individual and store specific factors. Our empirical analysis is based on the panel data from MarketingScan. We developed a Latent Factor Random Coefficients Multinomial Logistic Model that allows us to control for unobserved heterogeneity. We showed that the type of store brand had an influence on both the migration and the behavioural loyalty to store brand. The price is still important in the competition between private labels and national brands. Greater availability of SKUs at the brand level promotes migration to store brands, which helps them capture additional purchases. Nevertheless, a strong presence of private labels in a product category reduces consumer satisfaction with the assortment and consequently causes households to migrate to national brands.

nÃo hà / Esta dissertaÃÃo efetuou uma anÃlise empÃrica, baseada nas infraÃÃes de trÃnsito registradas no cadastro de veÃculos tipo motocicletas e similares de atà 125cc, ocorridas no Estado do CearÃ, no perÃodo de 2010 a 2013. O objetivo foi avaliar as caracterÃsticas sociodemogrÃficas que contribuem para aumentar a ocorrÃncia de reincidÃncia das infraÃÃes de trÃnsito por motociclistas, de acordo com as variÃveis: sexo, idade e tempo de habilitaÃÃo, e; a taxa de motorizaÃÃo e sua implicaÃÃo no fluxo de trÃnsito nos municÃpios cearenses. Para tanto, utilizouse o modelo economÃtrico Logit, cuja base de dados foi fornecida pelo Departamento Estadual de TrÃnsito â DETRAN-CE e pelo Instituto de Pesquisa e EstratÃgia EconÃmica do Cearà - IPECE. O que se observou, no Estado do CearÃ, à que os resultados corroboram as evidÃncias da literatura internacional: os infratores de trÃnsito reincidentes sÃo, em sua maioria, jovens e do sexo masculino, o que indica a necessidade de polÃticas pÃblicas de prevenÃÃo, incluindo esses grupos especÃficos da populaÃÃo. AlÃm disso, a condiÃÃo de habilitaÃÃo indica que os motociclistas mais experientes, ou com mais tempo de habilitaÃÃo, sÃo os que tÃm maior probabilidade de reincidÃncia. Para as estimaÃÃes acerca da taxa de motorizaÃÃo, os resultados sugerem que os condutores que estÃo nos municÃpios com taxas de motorizaÃÃo superiores à taxa de referÃncia de 20% apresentam uma maior probabilidade de reincidirem nas infraÃÃes de trÃnsito. / This paper made an empirical analysis based on traffic violations recorded in type vehicle registration motorcycles and the like up to 125cc, occurred in the state of CearÃ, in the period 2010 to 2013. The objective was to assess sociodemographic characteristics that contribute to increase the occurrence of repeated traffic violations by motorcyclists, according to variables such as gender, age and time of qualification, and; motorization rate and its implication in traffic flow in Cearà municipalities. For this, we used the logit econometric model, whose database has been provided by the State Department of Motor Vehicles â DETRAN-CE and the Institute of Economic Research and Strategy of Cearà - IPECE. What was observed in the state of CearÃ, is that the results support the evidence of the international literature: the repeat traffic offenders are mostly young and male, which indicates the need for preventive policies, including these specific population groups. In addition, the driver's condition indicates that the more experienced riders, or longer clearance, are the ones who are more likely to recur. For the estimations about the motorization rate, the results suggest that drivers who are in municipalities with motorization rates above the 20% reference rate are more likely to re-offend in traffic violations.