Search results

Abstract
The integration of multiple survey, administrative, and third-party data offers the opportunity to innovate and improve survey estimation via statistical modeling. With decreasing response rates and increasing interest for more timely and geographically detailed estimates, imputation methodology that combines multiple data sources to adjust for low unit response and allow for more detailed publication levels, including geographic estimates, is both timely and necessary. Motivated by the Advance Monthly Retail Trade Survey (MARTS) and Monthly Retail Trade Survey (MRTS), we propose Bayesian hierarchical multiple imputation-dependent data models with the goals of automating imputation for the MARTS by using historic MRTS data and providing geographically granular (state-level) estimates for the MRTS via mass imputation using third-party data and spatial dependence. As a natural byproduct of this approach, measures of uncertainty are provided. This article illustrates the advantages of applying established Bayesian hierarchical modeling techniques with multiple source data to address practical problems in official statistics and is, therefore, of independent interest. The motivating empirical studies are unified by their hierarchical modeling framework, which ultimately results in a more principled approach for estimation for the MARTS and a more geographically granular data product for the MRTS.

Intro -- MODELING MULTILEVEL DATA IN TRAFFIC SAFETY: A BAYESIAN HIERARCHICAL APPROACH -- MODELING MULTILEVEL DATA IN TRAFFIC SAFETY: A BAYESIAN HIERARCHICAL APPROACH -- Library of Congress Cataloging-in-Publication Data -- Contents -- Preface -- Chapter 1: Introduction -- Chapter 2: Review on Crash Prediction Models -- 2.1. Crash Frequency Prediction Models -- 2.1.1. Poisson Regression Model -- 2.1.2. Overdispersed-Poisson Regression Models -- 2.1.3. Zero-Inflated Count Model -- 2.2. Crash Severity Prediction Models -- 2.2.1. Binomial Logit and Probit Models -- 2.2.2. Multinomial Logit and Probit Models -- 2.2.3. Ordered Logit and Probit Models -- Chapter 3: Multilevel Data Structure in Traffic Safety -- 3.1. A Common Neglect in Existing Models: Between-Group Heterogeneity -- 3.2. A -level Hierarchy in Traffic Safety Data -- 1) Two-Entity-Level Design -- 2) Three-Entity-Level Design -- 3) Panel Data Design -- 4) Repeated Cross-Sectional Design -- Chapter 4: Bayesian Hierarchical Approach on Multilevel Crash Data -- 4.1. Hierarchical Model -- 1) Model Design -- 2) Model Estimation -- 3) Model Prediction -- 4.2. Bayesian Inference -- 4.3. Model Comparison and Assessment -- 4.3.1. Deviance Information Criterion -- 4.3.2. Cross Validation Predictive Densities -- Chapter 5: Illustrative Examples -- 5.1. Example 1: Crash Frequency Prediction at Intersections -- 5.1.1. Model Specification -- 5.1.2. Data Set for Analysis -- 5.1.3. Results and Discussions -- 5.1.4. Summary of Example 1 -- 5.2. Example 2: Severity of Driver Injury and Vehicle Damage in Intersection Crashes -- 5.2.1. Model Specification -- 5.2.2. Data Set for Analysis -- 5.2.3. Results and Discussions -- 5.2.4. Summary of Example 2 -- Chapter 6: Conclusion -- References -- Index.

Individuals who are more sensitive to negative outcomes from error are more likely to provide nonresponses in surveys. We argue Neurotics' sensitivity to negative outcomes leads them to avoid gathering costly information and forming/reporting opinions about stimuli. Using data from the 2014 Cooperative Congressional Election Study, we show Neuroticism is strongly and positively associated with NA/DK responses when placing politicians on a seven-point ideological scale. We then introduce to political science a Bayesian hierarchical model that allows nonresponse to be generated by both a lack of information as well as disincentives for response. Using this model, we show that the NA/DK responses in these data are due to inhibited information collection and indecision from error avoidance by Neurotics.

"This book is a welcome addition to the Bayesian literature. It is well written and amply illustrates Bayesian methods with practical applications in fisheries management. The programs for data analyses are available on the book's website, allowing users to get their 'hands dirty' and in the process really understand the model construction and the software."- Subhash R. Lele, Ecology, 95(1), 2014"The book is well written and easy to read, and the material presented deserves a greater exposure in taught statistics courses. I thoroughly recommend the book and believe that the statistical techniques and their application to quantitative fisheries science could ideally complement a short undergraduate course in applied statistics."-Carl M. O'Brien, International Statistical Review (2013), 81.

The Finnish salmonella control program (FSCP) for beef production is based on both randomized and selective testing of herds and animals. Sampling of individual animals in abattoirs is randomized. Herds are selectively tested for salmonella on the basis of clinical symptoms and/or other factors. Risk assessment of FSCP is inherently difficult due to the complexity of the complete data set, especially if the detailed labeling of the testing types is lost. However, for a risk assessment of the whole production chain, methods for exploiting all available data should be considered. For this purpose, a hierarchical Bayesian model of true salmonella prevalence was constructed to combine information at different levels of aggregation: herds in geographical regions and individual animals arriving for slaughter. The conditional (municipality specific) probability of selection of a herd for testing was modeled given the underlying true infection status of the herd and information about the general sampling activity in the specific region. The model also accounted for the overall sensitivity of the sampling methods, both at the herd and at the animal level. In 1999, the 95% posterior probability intervals of true salmonella prevalence in the herd population, in individual cattle, and in slaughter animal populations were [0.54%, 1.4%] (mode 0.8%), [0.15%, 0.39%] (mode 0.2%), and [0.12%, 0.36%] (mode 0.2%), respectively. The results will be further exploited in other risk assessments focusing on the subsequent parts of the beef production chain, and in evaluation of the FSCP.

Population estimates are critical for government services, development projects, and public health campaigns. Such data are typically obtained through a national population and housing census. However, population estimates can quickly become inaccurate in localized areas, particularly where migration or displacement has occurred. Some conflict-affected and resource-poor countries have not conducted a census in over 10 y. We developed a hierarchical Bayesian model to estimate population numbers in small areas based on enumeration data from sample areas and nationwide information about administrative boundaries, building locations, settlement types, and other factors related to population density. We demonstrated this model by estimating population sizes in every 10- m grid cell in Nigeria with national coverage. These gridded population estimates and areal population totals derived from them are accompanied by estimates of uncertainty based on Bayesian posterior probabilities. The model had an overall error rate of 67 people per hectare (mean of absolute residuals) or 43% (using scaled residuals) for predictions in out-of-sample survey areas (approximately 3 ha each), with increased precision expected for aggregated population totals in larger areas. This statistical approach represents a significant step toward estimating populations at high resolution with national coverage in the absence of a complete and recent census, while also providing reliable estimates of uncertainty to support informed decision making.

Population estimates are critical for government services, development projects, and public health campaigns. Such data are typically obtained through a national population and housing census. However, population estimates can quickly become inaccurate in localized areas, particularly where migration or displacement has occurred. Some conflict-affected and resource-poor countries have not conducted a census in over 10 y. We developed a hierarchical Bayesian model to estimate population numbers in small areas based on enumeration data from sample areas and nationwide information about administrative boundaries, building locations, settlement types, and other factors related to population density. We demonstrated this model by estimating population sizes in every 10- m grid cell in Nigeria with national coverage. These gridded population estimates and areal population totals derived from them are accompanied by estimates of uncertainty based on Bayesian posterior probabilities. The model had an overall error rate of 67 people per hectare (mean of absolute residuals) or 43% (using scaled residuals) for predictions in out-of-sample survey areas (approximately 3 ha each), with increased precision expected for aggregated population totals in larger areas. This statistical approach represents a significant step toward estimating populations at high resolution with national coverage in the absence of a complete and recent census, while also providing reliable estimates of uncertainty to support informed decision making.

Discarding is one of the most important topics in fisheries management, both for economic and ecological reasons. The European Union has included, through the current EU Common Fisheries Policy (CFP) Regulation, a discard ban with a quite controversial instrument: to enforce the landing of unwanted catch as a measure to promote their reduction. This management decision may condition the future of the fishing exploitation in European Sea. Within this context, both stakeholders and policy makers are now claiming for more effective tools that can be used to support the decision-making framework. In this study, we propose a simulation-based approach combining hierarchical Bayesian Spatial Models (H-BSMs) with the spatial-temporal module of Ecopath with Ecosim (EwE) approach, Ecospace, in the North Western Mediterranean Sea. In particular, we firstly assessed high-density discard areas using H-BSMs with fisheries and environmental data, and secondly, we simulated potential management options to identify the trade-offs of the discard ban application within these areas using EwE. We argue that coupling novel methods, as the ones used in this study, could be a decisive step to identify the best management action among a set of different scenarios within the context of the discard ban application in European Seas ; 1