Robust likelihood functions in Bayesian inference

In order to deal with mild deviations from the assumed parametric model, we propose a procedure for accounting for model uncertainty in the Bayesian framework. In particular, in the derivation of posterior distributions, we discuss the use of robust pseudo-likelihoods, which offer the advantage of preventing the effects caused by model misspecifications, i.e. when the underlying distribution lies in a neighborhood of the assumed model. The influence functions of posterior summaries, such as the posterior mean, are investigated as well as the asymptotic properties of robust posterior distributions. Although the use of a pseudo-likelihood cannot be considered orthodox in the Bayesian perspective, it is shown that, also through some illustrative examples, how a robust pseudo-likelihood, with the same asymptotic properties of a genuine likelihood, can be useful in the inferential process in order to prevent the effects caused by model misspecifications.     

Perceptions of intergenerational conflict: The politics of young vs. old in Florida

Not long ago, syndicated columnist Richard Reeves pondered the portents in an aging America and found trouble. “There is a confrontation coming in the United States between the demands of old people and the needs of the whole society,” he declared (Reeves 1988). His premonition is widely shared. Many social prophets are convinced that the graying of America is the prelude to a new politics. It will be the young against the old, they suggest, a politics of age polarization characterized by political mobilization, confrontation and backlash between generations reaching from Congress to city hall. Florida, with the nation's highest proportion of aging and a steady inflow of new retirees, is where many expect such conflict to appear early and often. This article concerns whether Florida is, or soon may be, the arena for such a confrontation and, if so, how it has become manifest.     

Equitable insurance premium schemes

We analyze the equity properties of insurance premium schemes where agents are partitioned into groups with different average accident probabilities and each individual has to pay a premium according to the average probability of the group to which it belongs. In particular, we examine the question whether choosing finer partitions to define these groups generates more equitable situations than coarser groups. Though it turns out that partitioning the agents into finer groups can never be Lorenz dominated by the coarser partition, it cannot be guaranteed that finer partitions represent improvements over coarser ones except in very restrictive circumstances. Received: 27 March 2000/Accepted: 28 August 2000     

Woman A buse in Dating Relationships: The Contribution of Male Peer Support*

The relationship between male peer support and premarital woman abuse is evaluated by integrating three bodies of knowledge, namely, woman abuse, social support, and reference others. Using data from a convenience sample of male undergraduate students, this study tests four hypotheses derived from a male peer support model. The findings support the model relating male peer group dynamics and woman abuse in university dating relationships.     

DO UNIONIZED FIRMS HIRE BETTER WORKERS?

It is a strong prior among many economists that unionized firms hire better-quality workers to offset higher union wages. In fact, standard economic theory does not support this prior. The key insights introduced by this paper are that, first, unions will likely raise future wages to reflect improvements in worker quality and, second, that unionized firms, anticipating this, often do better by hiring lower-quality workers. This surprising result has empirical support.     

Differentiation Levels of College Students: Effects on Vocational Identity and Career Decision Making

This study assessed the effects of differentiation levels on the career development of college students. Participants were 231 college students who completed the Differentiation of Self Inventory (Skowron & Friedlander, 1998 ), My Vocational Situation (Holland, Daiger, & Power, 1980 ), the Career Decision Profile (CDP; Jones & Lohmann, 1998 ), and demographic questions. The results supported the hypotheses that higher levels of the various components of differentiation would predict higher levels of vocational identity and fewer difficulties with career decision making. In particular, lower levels of emotional cutoff and emotional reactivity and higher levels of “I position” predicted higher levels of vocational identity and career decision making. Results suggest a more complicated picture for fusion, with higher levels being predictive of lower levels of decisiveness but not significantly related to vocational identity. Implications of the results for career counselors are provided.     

Truncated sequential sign tests for location shift

A truncated sequential sign test for location shift is studied when the null or target location has been estimated from a prior, fixed sample. If the randomness of the target is ignored then the test is shown to be strongly anticonservative, the degree being proportional to the ratio of the truncation point to the fixed sample size. The test is distribution-free under the hypothesis of no shift enabling exact Type I errors and null expected samples sizes to be calculated and compared to a modified Brownian motion approximation. A Monte Carlo power study shows that the test compares favorably with thr test against a Xnown target. An abbreviated table of critical values is given.