Analysis Of Event History Data Research Articles

Regression analysis of mixed panel count data with dependent observation processes

Event history data commonly occur in many areas and a great deal of literature on their analysis has been established. However, most of the existing methods apply only to a single type of event history data. Recently, several authors have discussed the analysis of mixed types of event history data and the existence of dependent observation processes is another issue that one often has to deal with in the analysis of event history data. This paper discusses regression analysis of mixed panel count data with dependent observation processes, which has not been addressed in the literature, and for the problem, an approximate likelihood estimation approach is proposed. For the implementation, an EM algorithm is developed and the proposed estimators are shown to be consistent and asymptotically normal. An extensive simulation study is performed to assess the performance of the proposed approach and indicates that it works well in practical situations. An application to a set of real data is provided.

Online community's recognition and continued participation in idea competitions

AbstractThis study examines the effect of online community's recognition on continued participation in idea competitions, and how personal winning record moderates such an influence. We reason that the motivating role of community recognition might either be reinforced or substituted by personal winning record, depending upon whether relational motives (psychological and social bonding) or individualistic motives (personal benefits, such as status and career enhancement) are the primary behavior driver. Through an event history analysis of data obtained from a platform of creative design contests, we find that although community recognition exerts a positive effect on the rate of continued participation for designers who are yet to win any competitions, this effect increasingly turns negative for designers who have won. Such findings indicate that the motivating role of community recognition might be substituted instead of reinforced by personal winning record, lending support to the individualistic view while rejecting the relational view. Although virtual social spaces represent an important means for modern competition platforms to attract and motivate participants, our study informs practitioners about online community's limitation in retaining “star” participants.

Additive hazard regression of event history studies with intermittently measured covariates

In the analysis of event history data, some covariates may vary over time. Such covariates are typically sparse and measured intermittently in many longitudinal settings. This article considers additive hazard regression models with intermittently measured covariates and proposes kernel weighting methods to estimate the model parameters. Two scenarios are investigated: half kernel estimation for data in which observation ceases at an event, and full kernel estimation in which observation may continue after an event, such as in recurrent event data. The asymptotic properties of the resulting estimators are established. Simulation studies show that the proposed estimators perform satisfactorily. An application to a study on primary biliary cirrhosis is illustrated.

SSI Participation Among the Elderly: A Hazard Model Approach

This study revisits the historic policy concern over nonparticipation in the Supplemental Security Income (SSI) program. The most contemporary research model suggests that from a cross-sectional perspective, the elderly take-up decision is primarily determined by the financial situation of the eligible individuals. Yet extant studies have yet to establish a longitudinal model of SSI participation that explicitly takes into account the changing circumstances of the elderly. As such, this study extends the current literature by exploring the time-varying processes leading to eventual take-up of benefits among the elderly. Methodologically, the research relies on event history analysis of data from the Health and Retirement Study spanning the years 1996 to 2006 to explain the varying rates of participation over time as well as the role of life events on the take-up decision.

A mixture of beta–Dirichlet processes prior for Bayesian analysis of event history data

In this paper, we propose a mixture of beta–Dirichlet processes as a nonparametric prior for the cumulative intensity functions of a Markov process. This family of priors is a natural extension of a mixture of Dirichlet processes or a mixture of beta processes which are devised to compromise advantages of parametric and nonparametric approaches. They give most of their prior mass to the small neighborhood of a specific parametric model. We show that a mixture of beta–Dirichlet processes prior is conjugate with Markov processes. Formulas for computing the posterior distribution are derived. Finally, results of analyzing credit history data are given.

A segmented regression model for event history data: an application to the fertility patterns in Italy

We propose a segmented discrete-time model for the analysis of event history data in demographic research. Through a unified regression framework, the model provides estimates of the effects of explanatory variables and jointly accommodates flexibly non-proportional differences via segmented relationships. The main appeal relies on ready availability of parameters, changepoints, and slopes, which may provide meaningful and intuitive information on the topic. Furthermore, specific linear constraints on the slopes may also be set to investigate particular patterns. We investigate the intervals between cohabitation and first childbirth and from first to second childbirth using individual data for Italian women from the Second National Survey on Fertility. The model provides insights into dramatic decrease of fertility experienced in Italy, in that it detects a ‘common’ tendency in delaying the onset of childbearing for the more recent cohorts and a ‘specific’ postponement strictly depending on the educational level and age at cohabitation.

Unemployment, Employment and Inactivity in Denmark: An Analysis of Event History Data

In this paper I estimate a discrete time hazard model for the exits from the different labour market states - unemployment, employment and inactivity (or OLF) - in the Danish labour market. I find that women and individuals over fifty are more likely to experience the long-term unemployment and inactivity. The less educated and unskilled workers are found to be another risk group to face the marginalisation from the labour market. Being previously employed reduces the risk of OLF, and increases the re-entry to employment probability, while living in the biggest Danish cities makes persons disadvantaged. These give the evidence that the Flexicurity model makes the weakest individuals disadvantaged in the Danish labour market. And finally, I find that those, who survived in a job one year, tend to remain employed, while persons, longer than one year inactive, face much higher risk of marginalisation.

Gender and Job Mobility in Postsocialist China: A Longitudinal Study of Job Changes in Six Coastal Cities

This study examines the gender differences in job mobility in urban China. Conceptualizing China's postsocialist transition as a multi-faceted process, we argue that the emergence of labor markets, gendered role differentiation within the family, and the state's declining involvement in promoting women's rights lead to widened gender gaps in job mobility. Event history analysis of data from six coastal cities finds that married women are less likely than their male counterparts to change jobs for career advancements, but are more likely to experience family-oriented job changes and involuntary terminations. The gender gaps in career-oriented job changes and involuntary terminations have also widened considerably during the reform period. We interpret these findings as the joint results of economic, socio-cultural and political processes.

On the inclusion of bivariate marked point processes in graphical models

Graphical models use graphs to represent conditional independence relationships among random variables of a multivariate probability distribution. This paper introduces a new kind of chain graph models in which nodes also represent marked point processes. This is relevant to the analysis of event history data, i.e. data consisting of random sequences of events or time durations of states. Survival analysis and duration models are particular cases. This article considers the case of two marked point processes. The idea consists of representing a whole process by a single node and a conditional independence statement by a lack of connection. We refer to the resulting models as graphical duration models.

BITE: A Bayesian Intensity Estimator

BITE is a software package designed for the analysis of event history data using flexible hierarchical models and Bayesian inference, with a particular emphasis on the application of flexible intensities as a description of the distribution of lifetimes. BITE provides a framework for combining flexible baseline hazard rates and observed data into intensity processes. Inclusion of covariate information is possible, and data can be non-informatively and independently filtered, or censored. The model and the data are described by a command language and data are stored into text files. Markov chain Monte Carlo methods are used for numerical approximation of expectations with respect to the posterior. Output consists of (i) parameter values stored during simulations, (ii) estimated expectations of functionals of parameters, or (iii) graphs (created with Splus or R software packages) presenting point-wise expectations (and credibility intervals) of the baseline hazard rates.

The importance of flexible hazard functions in the analysis of organizational survival data -- experiences from a cohort of Swedish commercial banks

The survival pattern of Swedish commercial banks during the period 1830--1990 is studied by parametric and non-parametric event-history methods. In particular we study the sensitivity of the conclusions reached with respect to the model used. It is found that the hazard is inversely U-shaped, which means that models that cannot allow for this type of hazard run into difficulties. Thus two of the most popular approaches in the analysis of event history data, the Gompertz and the Weibull models produce misleading results regarding the development of the death risk of banks over time. As regards the effect of explanatory variables on survival, on the other hand, most models are found to be robust and even in cases of misspecified baseline hazards, the estimated effects of the explanatory variables do not seem to be seriously wrong.

Tracing the Life Courses of Canadians

This paper examines the life courses of Canadians through an event history analysis of data from the 1990 General Social Survey on Family and Friends. The sequences and the timing of transitions into various life course stages, and the durations of stay in those stages are analyzed through multiple-decrement life tables. Changes in the life courses over 10-year birth cohorts from 1910 to 1970 and differentials by gender are highlighted in the paper. The analysis reveals that what is traditionally thought of as a typical life course is experienced only by one-fourth to one-third of a cohort. With the increasing diversity of life course stages among younger cohorts, it would no longer be adequate to consider only the typical life stages in future analyses.

Unfolding event histories: Scaling of non-repeating events

The present paper discusses the use of Coombs' (1950, 1964) unfolding procedure applied to event history data. This procedure is appropriate when several stimuli are rated according to the individuals' preference, and yields scale values for stimuli and individuals, as well as a ‘best’ order of stimuli. We propose to consider the order of occurrence in time of non-repeatable events (such as first marriage, leaving home, and first job) as an indication of the subjects' preference for these stimuli. In that case unfolding may be applied to the analysis of event history-data. The statistical principles of unfolding analysis in general are briefly discussed. Subsequently an illustration is given, using data concerning the life course of 216 Dutch women of birth cohort 1961, concerning five events: first steady relationship, leaving home, first child, first job, and first occurrence of living together. It is shown that these events can be ordered at an underlying one-dimensional continuum on which individuals can be scaled. Furthermore, using regression analysis it is shown that the location of the subjects at this continuum correlates significantly with various explanatory variables, showing how unfolding may be used in practical applications.

Long‐Term Effects of a Plant Closure: a Multistate Duration Analysis of Event History Data in a Ten‐Year Follow‐Up Study

AbstractThis paper discusses the possible effects on future prospects in the labour market of those who suffer a plant closure. A multistate duration model is estimated on monthly event data comprising 172 individuals over ten years, with the occurrence of the plant closure among the regressors. An individual can escape to one of several destinations when exiting one of three origin states (unemployed, sick leave, outside the labour force).Our prime concern has been transitions leading to a new job or to destination states outside the labour force (disability pension and at home, without any social security benefit). We find indications of severe negative long‐term effects on employment from a plant closure: long durations in unemployment when searching for new jobs, higher incidence of leaving the labour force, longer periods of sick leave and higher incidence of disability pension. Thus, an adverse shock reducing employment may have negative long‐term effects on the labour market through long spells of unemployment and a shrinking labour force. Our findings thus support the hypothesis of hysteresis dynamics in the labour market.

Interorganizational Linkages and Population Dynamics: Buffering and Transformational Shields

This research was partially supported by the University of Wisconsin Graduate Fund and the University of WisconsinMadison Business School Research Fund. Helpful suggestions during the paper's evolution were provided by William Barnett, Jack Brittain, Dan Levinthal, James March, Cora Marrett, Craig Olson, Jitendra Singh, and three anonymous reviewers. The order of authorship was determined randomly. This paper examines evidence concerning the existence of organizational buffers, which insulate an organization from environmental disturbances, and transformational shields, which insulate an organization from the risk of failure due to transformation. It proposes that interorganizational linkages in particular can (1) buffer organizations from failure, (2) affect the likelihood of organizational transformation, and (3) modify the effect of organizational transformation on failure. These propositions are examined using event-history analysis of data on approximately 1,000 Finnish newspapers over a 200-year period. In this population, linkages reduced failure, increased organizational transformations, and altered the chances for failure after transformation. More generally, the findings support the existence of organizational buffers and transformational shields, suggesting further work on their role in population dynamics.'

The analysis of event history data: a review of progress and outstanding problems.

This paper reviews methods for the analysis of event history data by both parametric (exponential/Poisson) and semi-parametric methods. It identifies a need for software for handling the data structures of complex event histories and shows that, with such software, most Markov and semi-Markov models of event history data may be dealt with in the framework of generalized linear models. Finally, the emergent 'frailty' models for associated risks are discussed together with their implications for statistical software.

The Transition to Adulthood

An enonnous amount of research on the transition to adulthood in America is founded on early landmark studies that include Elmtown' S Youth by August B. Hollingshead (1961), The Adolescent Society by James Coleman (1961), and the report of the Presidential Panel on Youth ( 1974). Research culminating in what has become known as the Wisconsin model of status attainment focused attention on the transition to adulthood; the study found that social psycholog­ ical and institutional factors operating in adolescence mediated the effects of socioeconomic origins on educational attainment, occupational attainment, and e arnings (Sewell & Hauser 1975). More recently, concern with high levels of premarital adolescent pregnancy among blacks and rising levels among whites prompted a great many studies on the antecedents and consequences of a premature transition to parenthood (Chilman 1980). Similarly, high levels of youth unemployment stimulated investigation of the process by which young people make the transition to full-time worker (Stephenson 1979). Scholarly interest in changes in the family life course has led to research on historical changes in the transition to adulthood (Modell et a11976, Winsborough 1978, Hogan 1981). Researchers' work on labor markets and socioeconomic attain­ ments led to study of the nature of careers (Spenner et a11982, Elder 1985a ). At the same time, new and better techniques for the collection and analysis of event history data have been developed (Hannan & Tuma 1979). The transition from adolescent to adult has long been of interest to students of human development. During the period of adolescence, young people grow to their full adult size , undergo the hormonal and physical changes associated with puberty, and attain reproductive maturity. Piaget has defined the cognitive task of adolescence as the achievement of fonnal operational reasoning (see Mussen 1970, Keating 1980). Personality development during adolescence has been

Social Dynamics: Models and Methods.

This book attempts to provide a broad framework for thinking about and conducting quantitative empirical studies of social change processes. The goals of this book are 1) to clarify and develop models and methods for causal analysis of dynamic social processes 2) to formulate continuous-time models of change in both quantitative and qualitative outcomes and to develop suitable methods for estimating these models from the kinds of data commonly available to sociologists 3) to develop a stochastic framework for analyzing both qualitative and quantitative outcomes and 4) to alter the way that sociologists think about the empirical study of social change processes. Part I provides general background for what follows; it includes both a discussion of the substantive importance of dynamic analyses in sociology and a review of models and methods previously used by sociologists interested in the empirical study of social dynamics. Part II contains 8 chapters on models and methods for anallyzing change in qualitative outcomes; it concentrates mainly on methods based on analysis of event-history data. Part III contains 6 chapters on comprable models and methods for analyzing change in quantitative outcomes; it focuses primarily on methods based on analysis of panel data.