Bayes methods in the ecological fallacy cntext:estimation of individual correlation from aggregate data

The ecological fallacy is related to Simpson's paradox (1951) where relationships among group means may be counterintuitive and substantially different from relationships within groups, where the groups are usually geographic entities such as census tracts. We consider the problem of estimating the correlation between two jointly normal random variables where only ecological data (group means) are available. Two empirical Bayes estimators and one fully Bayesian estimator are derived and compared with the usual ecological estimator, which is simply the Pearson correlation coefficient of the group sample means. We simulate the bias and mean squared error performance of these estimators, and also give an example employing a dataset where the individual level data are available for model checking. The results indicate superiority of the empirical Bayes estimators in a variety of practical situations where, though we lack individual level data, other relevant prior information is available.     

Nonparametric rank-based test procedures for non-additive models in the two-way layout I. no replications

One of the major unresolved problems in the area of nonparametric statistics is the need for satisfactory rank-based test procedures for non-additive models in the two-way layout, especially when there is only one observation on each combination of the levels of the experimental factors. In this paper we consider an arbitrary non-additive model for the two-way layout with n levels of each factor. We utilize both alignment and ranking of the data together with basic properties of Latin squares to develop rank tests for interaction (non-additivity). Our technique involves first aligning within one of the main effects, ranking within the other main effects (columns and rows) and then adding the resulting ranks within “interaction bands” corresponding to orthogonal partitions of the interaction for the model, as denoted by the letters of an n × n Latin square. A Friedman-type statistic is then computed on the resulting sums. This is repeated for each of (n?1) mutually orthogonal Latin squares (thus accounting for all the interaction degrees of freedom). The resulting (n?1) Friedman-type statistics are finally combined to obtain an overall test statistic. The necessary null distribution tables for applying the proposed test for non-additivity are presented and we discuss the results of a Monte Carlo simulation study of the relative powers of this new procedure and other (parametric and nonparametric) procedures designed to detect interaction in a two-way layout with one observation per cell.     

Caffeine consumption, expectancies of caffeine-enhanced performance, and caffeinism symptoms among university students

The expectancies paradigm developed in alcohol research was used to study caffeine consumption and signs of caffeinism in two groups of college students. A survey questionnaire was developed to gather self-report data on expectations of caffeine-enhanced performance (EP-CAFF), level of beverage caffeine consumed daily, and DSM-111 caffeinism signs reportedly experienced after consumption of caffeinated beverages. A positive association between EP-CAFF and both caffeine consumption and caffeinism signs was predicted. In both initial (n = 527) and cross-validation (n = 270) samples, significant correlations between EP-CAFF scores and both caffeine consumption and caffeinism symptoms were found. A small, but statistically significant, relationship was also found between caffeinism symptoms and both alcohol consumption and cigarette smoking. Of the 797 subjects in the combined sample, 151 (18.9%) endorsed five or more DSM-III caffeinism signs. Subjects endorsing five or more signs of caffeinism scored higher on the EP-CAFF scale, and consumed more caffeine and alcohol in both studies and smoked more cigarettes in study one--but not in study two--than did subjects who endorsed fewer than five caffeinism signs. Results support the utility of extending the expectancies model of substance use motivation from alcohol to caffeine.     

The influence of urine from asymptotic laboratory populations on sexual maturation in prairie deermice

Summary Male and female prairie deermice housed singly from weaning (21 days) until 40 days of age had urine from one of the following sources applied daily to their nostrils: sexually mature laboratory population males or females, and sexually mature breeder males or females from our laboratory colony. Other males and females were treated with water. Records were obtained of the weights of the tests, seminal vesicles and bacula of males and the ovaries and uteri of females at 40 days of age. The major results of this study were as follow: Female urine from sexually mature and inhibited population or isolated mice stimulated the development of the male reproductive organs significantly more than water or male urine. Male urine applied to young males did not differentially affect the weights of their reproductive organs compared with water and thus showed no inhibition effects. The reproductive organs of females treated with male urine did not differ significantly in weight from those of females treated with female urine. Females treated with water developed significantly larger uteri than those treated with urine from sexually mature population females (P≤0.02), inhibited population females (P<0.001) and inhibited population males (P<0.007).     

Illness career descent in institutions for the elderly

This paper examines the social psychological consequences of illness careers in a retirement facility. As elderly residents' health declines, they experience an illness career descent. This refers to the movement through the social structure of the healthcare facility as a consequence of the downward trajectory of chronic illness. Illness career descent requires the formal relocation of the individual and the informal social consequences of loss of control, stigmatization, and segregation from prior friendship networks. Individuals react in a three-stage manner. At first, residents resist being moved down. Next, there is a negative impact on their self-image as relocated residents accept the evaulation by staff that they have become less competent. Third, they adjust to their new location and enhance their self-image by developing new roles, finding a poor dear, re-evaluating prior stereotypes, and engaging in post-decisional dissonance reduction.The author wishes to express his deepest appreciation to Jacob Climo, Peter Conrad, Stan Kaplowitz, and Shulamit Reinharz for their encouragement and their suggestions on earlier drafts of this paper.     

Making Privatization Work: Utilizing a Scorecard Model of Human Resource Strategy

Three-fourths of privatization efforts fail. Analyses of the other one-fourth have overlooked what is arguably the single most important element: successful management of human resources. In establishing the culture of change necessary to effect competitive initiatives and in motivating employees toward innovation and accountability, a scorecard-based human resource strategy is at the center of it all. A discussion of the strategic role of human resource management supporting privatization efforts in general, and of the remarkably successful Charlotte, North Carolina experience in particular, offers important lessons. Detailed here is a rubric of what works and why.     

The Potential Impact of ‘Public Access’ Legislation on Access to Forestry Literature

There have been increasing calls for the United States (U.S.) government’s implementation of broad public access policies mandating free online access to federally funded research. This study examines the potential impact of such a policy on peer-reviewed forestry literature. The authors analyze information about federal government authorship, federal government funding, and U.S. authorship indicated in articles published in five core forestry journals in 2006. The results of the analysis provide evidence that federal public access legislation would have a significant impact on the accessibility of forestry literature published in leading journals in the field.     

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.     

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.