Influence analysis of structural equation models with polytomous variables

The estimation of model parameters in structural equation models with polytomous variables can be handled by several computationally efficient procedures. However, sensitivity or influence analysis of the model is not well studied. We demonstrate that the existing influence analysis methods for contingency tables or for normal theory structural equation models cannot be applied directly to structural equation models with polytomous variables; and we develop appropriate procedures based on the local influence approach of Cook (1986). The proposed procedures are computationally efficient, the necessary bits of the proposed diagnostic measures are readily available following an usual fit of the model. We consider the influence of an individual cell frequency with respect to three cases: when all parameters in an unstructured model are of interest, when the unstructured polychoric correlations are of interest, and when the structural parameters are of interest. We also consider the sensitivity of the parameters estimates. Two examples based on real data are presented for illustration.The work described in this paper was partially supported by a Chinese University of Hong Kong Postdoctoral Fellows Scheme and a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (RGC Ref. No. CUHK4186/98P). We are indebted to P.M. Bentler and M.D. Newcomb for providing the data set, Michael Leung for his assistance, and the Editor and the referees for some very valuable comments.     

School Effectiveness Findings 1979–2002

The London studies of school effectiveness in the 1970s provided evidence that the qualities of schools made a significant difference to pupil progress, after taking account of the characteristics and backgrounds of the pupils at the time of school entry. The purpose of this article is to review the further evidence that has accumulated since then in order to assess how far the findings have been confirmed or refuted, what new questions have arisen and what key challenges remain. The continuing areas of research and debate include school effects on behavioral outcomes, the effects of class size, the importance of parental involvement, the impact of intake mix, peer influences, and level of resources. There has been greater agreement on the many features fostering school effectiveness. However, questions remain on how to bring about desired changes. The paper concludes by drawing attention to 10 largely unanswered questions that matter for policy and practice.     

Review

Book reviewed in this article:
David Ray Griffin, Religion and Scientific Naturalism: Overcoming the Conflicts     

God and Evolutionary Evil: Theodicy in the Light of Darwinism

Pain, suffering, death, and extinction have been intrinsic to the process of evolution by natural selection. This leads to a real problem of evolutionary theodicy, little addressed up to now in Christian theologies of creation. The problem has ontological, teleological, and soteriological aspects. The recent literature contains efforts to dismiss, disregard, or reframe the problem. The radical proposal that God has no long–term goals for creation, but merely keeps company with its unfolding, is one way forward. An alternative strategy to tackle the problem of evolutionary theodicy is outlined, with an implication for environmental ethics and suggestions for further work.     

New Evolutionary Perspectives on Altruism: Multilevel-Selection and Costly-Signaling Theories

Social psychological theories tend to be primarily concerned with the immediate causes of altruism, whereas evolutionary explanations focus more on the origins and ultimate functions of altruistic behavior. Recent developments in the evolutionary psychology of altruism promise an even richer understanding of this important category of social behaviors. Specifically, new perspectives offered by multilevel-selection theory and costly-signaling theory may help to shed light on some of the more problematic issues in the study of altruism.     

实验设计:一种非模式的方法

实验设计通常被当作一种种设计模式．这种方法可称之为实验设计的模式方法。但本文认为,这种常用的方法事实上存在着不少缺点。于此．本文尝试提出了一种新的实验设计方法——实验设计的变量方法。实验设计的变量方法的基本思想是．从实验研究的问题出发,根据实验中变量的性质或特点确定实验设计的具体内容。本文最后讨论了实验设计变量方法的优越性。     

潜变量交互效应分析方法

简要回顾了分析显变量交互效应的常用方法。详细讨论了目前分析潜变量交互效应的主要方法,包括用潜变量的因子得分做回归分析、分组线性结构方程模型分析、加入乘积项的结构方程模型分析和两步最小二乘回归分析,并比较和评价了这些方法的优缺点。最后归纳了潜变量交互效应分析方法的研究趋势,并介绍了新近进展(包括LMS方法和GAPI方法)。     

A Biodata Inventory Administered via Interactive Voice Response (IVR) Technology: Predictive Validity, Utility, and Subgroup Differences

Although organizations are increasingly using interactive voice response (IVR) systems to help screen and select job applicants, few if any published studies have investigated the effectiveness of such systems. The current study responded to this need by evaluating a biodata inventory administered via IVR technology. A predictive validity study revealed that both a 42-item biodata measure and a more parsimonious 6-item composite correlated with supervisor ratings of performance and did not result in adverse impact or differential prediction. In addition, a utility analysis indicated that both IVR biodata measures would likely result in considerable financial gains to the hiring organization. The implications of these results for selection research and practice are discussed.     

近十年国外罗夏测验攻击变量研究综述

用罗夏测验测量攻击有助于了解个体的攻击性冲动 ,也有助于外显攻击行为的预测和控制。近十年国外文献中有关研究者主要有Holt、Exner、Gacono和Meloy。综述结果表明 ,他们提出的大多数攻击变量都有明确的定义 ,其实证研究也获得了一些支持。但也可以看到 ,目前的研究依然存在不少问题 :(1 )变量的定义有重复、内容有交叉 ;(2 )变量的分类标准不唯一 ;(3 )某些变量的使用和操作欠方便。因而 ,这些变量的定义、分类和使用都还有待于进一步研究支持。     

The capacity to be an analyst: A contribution from attachment research to the study of candidate selection

In this paper the author discusses how the study of candidate selection, once a topic of vibrant research, has unfortunately languished. Certain qualities were thought to characterize the successful candidate. However, they were never successfully operationalized nor empirically tested. Possibly because of this lack of empirical data, selectors today have difficulty articulating their criteria and are relying on intuition. In order to provide a more rational basis for contemporary selection, the author looks to the attachment literature. This makes sense because attachment theory shares some basic assumptions of contemporary psychoanalysis. The Adult Attachment Interview (AAI) is a research tool that predicts the ability of a parent to convey attachment security. It is scored by attending to how a person speaks about his early attachment experiences. The AAI appears to tap into similar qualities to those selection researchers have sought in their candidates. Further, the scoring method of the AAI appears to be similar to the last attempt by selection researchers to operationalize them. Given these similarities, the author recommends an empirical study using the AAI to operationalize these qualities in analytic candidates. The study would test their importance for success in the training program, thus offering selectors some empirical grounding for their choices.