Discrete-time Proportional Hazards Models Research Articles

Does a “survival-by-exporting” effect for SMEs exist?

The trade literature has long discussed the existence of some benefits attributed to exporting (learning-by-exporting), among others, the improvement in survival chances. This paper examines whether exporting SMEs enjoy better survival prospects than non-exporting SMEs. We investigate the determinants of survival of exporting and non-exporting SMEs and explore whether the exporting behaviour plays a significant role in explaining their probability of exit. For this purpose, we estimate discrete time proportional hazard models that account for unobserved individual heterogeneity. The dataset is a sample of Spanish manufacturing SMEs drawn from the Encuesta sobre Estrategias Empresariales (ESEE) for 1990–2002. After controlling for firm, industry and economy characteristics, we find evidence supporting the existence of a sizeable “surviving-by-exporting” effect. That is, exporting SMEs face a significantly lower probability of failure than non-exporters.

Factores socioeconómicos y contextuales que determinan la actividad reproductiva de las adolescentes en Colombia

To contribute to a better understanding of the problems of pregnancy among adolescent women, including proximate and socioeconomic determining factors, in two large, culturally different cities in Colombia: Santa Fe de Bogotá and Cali. This longitudinal study combined quantitative and qualitative research methods, using information generated by a survey of adolescents conducted in 2003. The survey included 550 adolescents in Bogotá and 550 adolescents in Cali, from all socioeconomic strata. To analyze the determinants, discrete-time proportional hazards models were used. For the qualitative study, 72 in-depth interviews and four focus groups were done. With the information organized by subjects and categories that were defined in relation to the purposes of the study, categories were identified that arose from the patterns and recurrences in the data, in order to see sociocultural trends by sex, stratum, and city. The patterns of sexual activity, union (married or unmarried relationship), and maternity differ considerably among the socioeconomic strata, in both of the cities. The adolescent women in the low stratum begin having sexual relations, form unions, and become mothers earlier in life and with greater frequency than do adolescent women in the medium or high strata. The main determinant of the reproductive behavior of adolescent women is the set of contextual and socioeconomic factors in the home, mainly the family context (environment and supervision) and the educational climate (the average number of years of formal education of the family members over the age of 15). Sex education has been provided in the schools in Colombia since 1993, but our results clearly indicate that it has had only a limited impact on the reproductive behavior of adolescent women.

Traumatic Experiences in Childhood and the Risk of Arthritis

Background: Recent data suggest that psychosocial factors, including childhood and adulthood stressors, may play a significant role in the development of chronic musculoskeletal pain and other symptoms. The purpose of this study was to determine if traumatic experiences in childhood are associated with an increased risk of self-reported arthritis later in life. Methods: We used longitudinal data (N=9,159) from the first 3 cycles of the National Population Health Survey (NPHS) in Canada. New cases of arthritis were identified using an interviewer-administered questionnaire. Psychological trauma in childhood or adolescence was measured by a 7-item questionnaire asking about physical abuse, fearful experiences, hospitalization, being sent away from home, and 3 types of parental disturbance. The effects of trauma were examined in a multivariable discrete-time proportional hazards model. Results: The incidence of self-reported arthritis was 27.1 per 1,000 person-years. We found a relative risk of 1.17 (95% CI=0.92, 1.48) for one traumatic event and 1.27 (95% CI=0.99, 1.62) for two or more traumatic events. Independent effects were observed for prolonged hospitalization (HR=1.33, 95% CI=1.05, 1.68) and being very scared (HR=1.29, 95% CI=1.02, 1.62). In subgroup analyses, no significant interactions were found between trauma and sex, socio-economic status, or baseline health. Conclusions: In this large prospective study, we found a moderate increase in the risk of arthritis among persons reporting multiple traumatic experiences in childhood.

The Proportional Hazard Model for Purchase Timing

We use the proportional hazard model (PHM) to study purchase-timing behavior of households in two product categories: laundry detergents and paper towels. The PHM decomposes a household's instantaneous probability of buying the product at a point of time into two components: the baseline hazard that captures the household's intrinsic purchase pattern over time and the covariate function that captures the effects of marketing variables on the household's purchase timing decision. We compare the continuoustime and discrete-time PHMs, where the latter explicitly accounts for households' shopping trips that do not involve purchase of the product. We find that the discrete-time PHM empirically outperforms the continuous-time PHM in terms of explaining the observed purchase outcomes. We compare five different parametric specifications of the baseline hazard, and find that the three-parameter expo-power specification outperforms the exponential, Erlang-2, Weibull, and log-logistic specifications. We use a causespecific, competing-risks PHM to distinguish between two types of purchase events that differ in terms of whether or not they were preceded by a shopping trip that involved purchase of the product. Such a cause-specific, competing-risks PHM is shown to outperform the traditional discrete-time PHM. We then estimate a nonparametric version of the PHM and find that it does not offer any additional insights compared to the parsimonious parametric PHM. Finally, we accommodate unobserved heterogeneity across households by allowing all of the parameters of the PHM to follow a discrete distribution across households whose locations and supports are nonparametrically estimated from the data. We find evidence for substantial unobserved heterogeneity in the data, both in the parameters of marketing variables and in the baseline hazards. This study will be a useful reference to researchers hoping to use the PHM to study event times.

A Monte Carlo EM algorithm for random-coefficient-based dropout models

Longitudinal studies of neurological disorders suffer almost inevitably from non-compliance, which is likely to be non-ignorable. It is important in these cases to model the response variable and the dropout mechanism jointly. In this article we propose a Monte Carlo version of the EM algorithm that can be used to fit random-coefficient-based dropout models. A linear mixed model is assumed for the response variable and a discrete-time proportional hazards model for the dropout mechanism; these share a common set of random coefficients. The ideas are illustrated using data from a five-year trial assessing the efficacy of two drugs in the treatment of patients in the early stages of Parkinson's disease.

Are survival estimates and models obtained by capture-recapture correct? A triple analysis of data with known dates of death

We use data from a survey of women with metastatic breast cancer coming more or less regularly for check-ups as a real-size test of capture-recapture methodology and models. Unlike capture-recapture data of animals, the date of death of each individual is exactly known. The individuals were separated into four groups; the study lasted for 10 years separated into 32 visit occasions (one every fourth month). We analysed the data three times: once with standard survival analysis with known date of death (discrete-time proportional hazards models); once with capture-recapture models after removing the date of death (using 'live recaptur' records); and once with 'return rates' corrected by later encounters of the individuals ('minimum number alive' estimates). Comparing the results obtained from the three methods confirms the efficiency of capture-recapture analysis.

Functional Health and Dental Service Use Among Older Adults Teresa Adults

Although socioeconomic barriers to receiving adequate dental care have been well documented, physical frailty as a risk factor for not visiting the dentist has not been fully explored. This study prospectively examines the relationship between functional health and dental service use, taking into account sociodemographic characteristics, general and dental health status, and prior dental utilization behavior. Data from a randomized trial of a comprehensive geriatric assessment and prevention program in community-dwelling adults age 75+ years living in Santa Monica, CA, collected between 1988 and 1993, were analyzed. A series of discrete-time proportional hazards models were used to assess the effects of functional status, sociodemographic characteristics, and general health and dental health measures on dental service use. Functional status was negatively associated with dental service use, and the conditional probability of a first visit to the dentist after baseline decreased over time. When additional measures of general health, dental health, and socioeconomic status were introduced, the effect of functional status was mitigated but remained significant. In the most fully specified model, which took dental visitation behavior prior to the beginning of the study into account, the effect of functional limitation remained significant. Even in this relatively well-educated group of older persons with higher than average dental service use, impaired functional status was associated with lower levels of dental service use over time.

Modelling Heaping Effects in Unemployment Duration Models – With an Application to Retrospective Event Data in the German Socio- Economic Panel / Ökonometrische Modellierung von „Heaping“-Effekten in Verweildauermodellen der Arbeitslosigkeit auf der Basis der retrospektiven Kalendariendaten des Sozio-ökonomischen Panels für Deutschland

Summary Duration data derived from retrospective surveys often show an abnormal concentration of responses at certain durations. This common kind of measurement error is known as “heaping” in the statistical literature. We show how heaping effects can be modelled within a maximum likelihood framework using external validation information and demonstrate how parameter estimates in discrete-time proportional hazard models are affected by alternative specifications of the heaping mechanism. The model is applied to unemployment duration data derived from the retrospective calendar information in the German Socio-Economic Panel. Our main result is that parameter estimates are generally rather insensitive to whether or not heaping is explicitly taken into account and to different assumptions about the heaping mechanism.

Linking Household Transitions and Housing Transitions: A Longitudinal Analysis of Renters

In this paper I use discrete-time proportional hazards models of competing risks to assess the association between household transitions and housing transitions for individuals in the rental sector. Specifically, moves within the rental sector are treated as a competing risk with moves to homeownership. A series of longitudinal models indicates the differential role household transition serves in predicting each of these movement types. This research confirms the association between relative household stability, relative income level, and the move to homeownership. Conversely, homeownership remains unattainable for individuals with relatively low incomes, in relatively transient household types, and especially for minorities.

Female Employment and First Union Dissolution in Puerto Rico

We examine the impact of employment histories on union dissolution in Puerto Rico for two groups of women: all women either married or living with a partner consensually and women who did not cohabit before legal marriage. Results using discrete-time proportional hazards models indicate that: (a) transitions in female employment status within a union increase the risk of union disruption; (b) transitions in female employment status relative to union inception increase the risk of union dissolution but this effect disappears when employment transitions within the union are controlled for; (c) working in jobs that pay a wage or salary increases union disruption as compared to women who do not work and women who work in more home-oriented occupations; and (d) working in types of occupations that are conducive to home-oriented roles does not affect the likelihood of union disruption. (EXCERPT)

Parametric and Parametrically Smoothed Distribution-Free Proportional Hazard Models with Discrete Data

This paper discusses discrete time proportional hazard models and suggests a new class of flexible hazard functions. Explicitly modeling the discreteness of data is important since standard continuous models are biased; allowing for flexibility in the hazard estimation is desirable since strong parametric restrictions are likely to be similarly misleading. Simulation compare continuous and discrete models when data are generated by grouping and demonstrate that simple approximations recover underlying hazards well and outperform nonparametric maximum likelihood estimates in term of mean squared error.