Improved Transferability of Data‐Driven Damage Models Through Sample Selection Bias Correction

Damage models for natural hazards are used for decision making on reducing and transferring risk. The damage estimates from these models depend on many variables and their complex sometimes nonlinear relationships with the damage. In recent years, data‐driven modeling techniques have been used to capture those relationships. The available data to build such models are often limited. Therefore, in practice it is usually necessary to transfer models to a different context. In this article, we show that this implies the samples used to build the model are often not fully representative for the situation where they need to be applied on, which leads to a “sample selection bias.” In this article, we enhance data‐driven damage models by applying methods, not previously applied to damage modeling, to correct for this bias before the machine learning (ML) models are trained. We demonstrate this with case studies on flooding in Europe, and typhoon wind damage in the Philippines. Two sample selection bias correction methods from the ML literature are applied and one of these methods is also adjusted to our problem. These three methods are combined with stochastic generation of synthetic damage data. We demonstrate that for both case studies, the sample selection bias correction techniques reduce model errors, especially for the mean bias error this reduction can be larger than 30%. The novel combination with stochastic data generation seems to enhance these techniques. This shows that sample selection bias correction methods are beneficial for damage model transfer.     

Multiple event times in the presence of informative censoring: modeling and analysis by copulas

Motivated by a breast cancer research program, this paper is concerned with the joint survivor function of multiple event times when their observations are subject to informative censoring caused by a terminating event. We formulate the correlation of the multiple event times together with the time to the terminating event by an Archimedean copula to account for the informative censoring. Adapting the widely used two-stage procedure under a copula model, we propose an easy-to-implement pseudo-likelihood based procedure for estimating the model parameters. The approach yields a new estimator for the marginal distribution of a single event time with semicompeting-risks data. We conduct both asymptotics and simulation studies to examine the proposed approach in consistency, efficiency, and robustness. Data from the breast cancer program are employed to illustrate this research.

     

Global Responses to the COVID-19 Pandemic

Public Organization Review -     

Does a universal non-contributory social pension make sense for rural China?

China’s pension reform during the past three decades has allowed a majority of China’s population to be covered by a pension scheme. Of particular note has been the New Rural Pension Scheme (NRPS), a voluntary programme introduced starting in 2009. One goal of our analysis is to assess that pension scheme, using a variety of sources of information including data drawn from recent (2013 and 2015) nationwide China Health and Retirement Longitudinal Surveys (CHARLS). Our analysis involves an exploration of differences between the generosity and structure of the NRPS and other pension schemes currently in place. We also explore the feasibility of reforming the current “quasi-social pension” component of the NRPS by substituting a universal non-contributory social pension pillar. In connection with our assessment of the NRPS, we note the unusually low benefit levels for rural China.     

Media Disaster Reporting Effects on Public Risk Perception and Response to Escalating Tornado Warnings: A Natural Experiment

Previous research has evaluated public risk perception and response to a natural hazards in various settings; however, most of these studies were conducted either with a single scenario or after a natural disaster struck. To better understand the dynamic relationships among affect, risk perception, and behavioral intentions related to natural disasters, the current study implements a simulation scenario with escalating weather intensity, and includes a natural experiment allowing comparison of public response before and after a severe tornado event with extensive coverage by the national media. The current study also manipulated the display of warning information, and investigated whether the warning system display format influences public response. Results indicate that (1) affect, risk perception, and behavioral intention escalated as weather conditions deteriorated, (2) responses at previous stages predicted responses at subsequent stages of storm progression, and (3) negative affect predicted risk perception. Moreover, risk perception and behavioral intention were heightened after exposure to the media coverage of an actual tornado disaster. However, the display format manipulation did not influence behavioral responses. The current study provides insight regarding public perception of predisaster warnings and the influence of exposure to media coverage of an actual disaster event.     

Strategic Alliance Research in the Era of Digital Transformation: Perspectives on Future Research

The emerging digital transformation in the twenty‐first century is rapidly and significantly changing the business landscape. The fast‐changing activities, expectations and new modes of collaboration suggest it is time to review the current theoretical insights from strategic alliance (SA) research, which are based on assumptions from a different era. We therefore aim to stimulate multidisciplinary debate and theoretical reflections to better understand emerging paradoxes and challenges that contemporary firms face in the formation, evolution and dissolution of strategic alliances. Specifically, we offer alternative visions of SA research and suggest fresh applications or supplements of existing theoretical perspectives and research methods that can better address the research questions emerging from an era of digital transformation.     

Spatial Variations in Church Burnings: The Social Ecology of Victimized Communities in the South*

Abstract In recent years, church burnings in the South have attracted a great deal of attention. Many commentators have charged that they are a product of strained race relations throughout the South, and particularly of severe racial tensions in Southern rural areas. In this study we evaluate these claims. We begin by mapping the spatial coordinates of recorded church burnings from 1990 to 1997, and find that church arsons indeed are concentrated in the South. Church burnings, however, are a more urban phenomenon than popular media accounts would suggest. Our analysis then explores the influence of contextual factors (population and locale, racial composition and inequality, so‐cioeconomic conditions, local religious ecology, and patterns of reported crime) on church burnings in counties located in the study region. Logistic regression models confirm that church arsons are most likely to occur in small metropolitan statistical areas (MSAs) and non‐MSA counties containing a city of at least 10,000 residents. Church burnings also are especially likely in counties with a higher percentage of black residents, a larger number of churches relative to the rest of the state, and a higher arson rate. We conclude by discussing the implications of these findings for future research and public policy.