Maximum likelihood from spatial random effects models via the stochastic approximation expectation maximization algorithm

We introduce a class of spatial random effects models that have Markov random fields (MRF) as latent processes. Calculating the maximum likelihood estimates of unknown parameters in SREs is extremely difficult, because the normalizing factors of MRFs and additional integrations from unobserved random effects are computationally prohibitive. We propose a stochastic approximation expectation-maximization (SAEM) algorithm to maximize the likelihood functions of spatial random effects models. The SAEM algorithm integrates recent improvements in stochastic approximation algorithms; it also includes components of the Newton-Raphson algorithm and the expectation-maximization (EM) gradient algorithm. The convergence of the SAEM algorithm is guaranteed under some mild conditions. We apply the SAEM algorithm to three examples that are representative of real-world applications: a state space model, a noisy Ising model, and segmenting magnetic resonance images (MRI) of the human brain. The SAEM algorithm gives satisfactory results in finding the maximum likelihood estimate of spatial random effects models in each of these instances.     

Lifetime Buy Decisions with Multiple Obsolete Parts

Life‐cycle mismatch occurs when the life cycles of parts end before the life cycles of the products in which those parts are used. Lifetime buys are one tactic for mitigating the effect of part obsolescence, where a quantity of parts is purchased for the remaining life of a product. We extend prior work that determines optimal lifetime buy quantities for one product with one obsolete part by providing an analytic solution and two simple heuristic policies for the optimal lifetime buy quantities when many parts become obsolete over a product's life cycle. We determine which of our two heuristics is most accurate for different product life cycles, which yields a metaheuristic with increased accuracy. That analysis also reveals critical perspectives in making lifetime buy decisions with nonstationary life‐cycle demand patterns.     

Product Design for Life‐Cycle Mismatch

Life‐cycle mismatch occurs when the life cycle of a product does not coincide with the life cycles of the parts used in that product. This is particularly a problem with products that contain electronic components that sometimes have life spans of only two years. The cost of mitigating component obsolescence, which may require redesigning the product, is often considerable. Thus, prudent product design necessitates the selection of electronic components and product architecture, considering the cost of mitigating an obsolete design and other costs related to the design and manufacture of a product. Accordingly, we develop and analyze a model that shows how a product design can be effectively tailored to a particular product's life cycle.     

Bayesian D-optimal supersaturated designs

We introduce a new class of supersaturated designs using Bayesian D-optimality. The designs generated using this approach can have arbitrary sample sizes, can have any number of blocks of any size, and can incorporate categorical factors with more than two levels. In side by side diagnostic comparisons based on the E(s²)$E(s^2)$ criterion for two-level experiments having even sample size, our designs either match or out-perform the best designs published to date. The generality of the method is illustrated with quality improvement experiment with 15 runs and 20 factors in 3 blocks.     

Multivariate Parametric Spatiotemporal Models for County Level Breast Cancer Survival Data

In clustered survival settings where the clusters correspond to geographic regions, biostatisticians are increasingly turning to models with spatially distributed random effects. These models begin with spatially oriented frailty terms, but may also include further region-level terms in the parametrization of the baseline hazards or various covariate effects (as in a spatially-varying coefficients model). In this paper, we propose a multivariate conditionally autoregressive (MCAR) model as a mixing distribution for these random effects, as a way of capturing correlation across both the regions and the elements of the random effect vector for any particular region. We then extend this model to permit analysis of temporal cohort effects, where we use the term temporal cohort to mean a group of subjects all of whom were diagnosed with the disease of interest (and thus, entered the study) during the same time period (say, calendar year). We show how our spatiotemporal model may be efficiently fit in a hierarchical Bayesian framework implemented using Markov chain Monte Carlo (MCMC) computational techniques. We illustrate our approach in the context of county-level breast cancer data from 22 annual cohorts of women living in the state of Iowa, as recorded by the Surveillance, Epidemiology, and End Results (SEER) database. Hierarchical model comparison using the Deviance Information Criterion (DIC), as well as maps of the fitted county-level effects, reveal the benefit of our approach.     

Structural equation modeling of drinking motives in male DUI offenders

Measures completed by 250 men convicted of driving under the influence of alcohol (DUI) were analyzed to examine the direct effects of parental problem drinking and the mediational influence of drinking motives on alcohol use and drinking-related problem outcomes. Structural equation modeling was used to evaluate competing models in the cross-sectional data set. Results indicated that: 1) reported parental problem drinking modestly predicted both alcohol use and problem outcomes when tested in the absence of influence from drinking motives; 2) confirmatory factor analysis yielded a 4-factor drinking motives solution consisting of social, enhancement, coping with negative emotions, and interpersonal facilitation; and 3) both enhancement and interpersonal facilitation drinking motives mediated parental problem drinking's direct effect on alcohol use and problem outcomes. Drinking to cope with negative emotions, while not mediational in nature, produced a direct effect pathway to problem outcomes. The findings extend predictive utility of specific drinking motive dimensions on alcohol use and problem outcomes from previous research in community and university samples to men convicted of driving while intoxicated.     

A Second Look at Self‐employment and the Earnings of Immigrants1

Portes and Zhou (1996) present evidence to suggest that self‐employment may impact the economic fortunes of immigrant groups by generating a relatively large number of very high earners. The present analysis seeks to extend and clarify the very high earner thesis using 1990 U.S. census data. Findings across 41 immigrant samples are generally consistent with expectations. However, in the majority of cases the effect of self‐employment is due in whole or substantial part to the earnings of professional specialists. Findings are consistent with a very high earner model of nonprofessional specialist self‐employment in eleven immigrant subsamples.     

Sexual assault prevention with college-aged women: a bibliotherapy approach

The present research evaluated the efficacy of a skills-based bibliotherapy approach to sexual assault prevention for college-aged women. One hundred and ten participants were followed prospectively for 16 weeks. A self-help book, written by the authors, was compared to a wait-list control on several self-report measures. Results revealed significant differences between groups, with bibliotherapy participants reporting decreased participation in risky dating behaviors and improvement in sexual communication strategies across a variety of dating situations. However, results suggested that the self-help book was no more effective than the wait-list control in reducing rates of sexual victimization. Limitations of the study and directions for future sexual assault prevention research with women are discussed.     

BECAUSE IT'S FREE POKER! A QUALITATIVE ANALYSIS OF FREE TEXAS HOLD'EM POKER TOURNAMENTS

The game of poker has dramatically increased in popularity in America over the past decade, and includes a new trend in poker in which tournaments are played with no monetary wagering. These “free” poker tournaments are the primary focus of this project. Our interest in these free poker tournaments was twofold: first, we wished to identify what would motivate a person to play poker when there is no profit motive; second, we wanted to learn how the social dynamics of the game itself may or may not change when monetary risk is not present. Through covert participant observation of free poker tournaments at four different locations we were able to discover three main motivations among the players: sharpening skills for later cash games, sociability, and gaining the status of a “regular player.” We also discovered that the tenor and tempo of the game tended to vary not as a function of the location, but rather because of the social class of the players. We offer several hypotheses to guide future research endeavors in this area.