Using Two-Level Factor Analysis to Test for Cluster Bias in Ordinal Data

The test for cluster bias is a test of measurement invariance across clusters in 2-level data. This article examines the true positive rates (empirical power) and false positive rates of the test for cluster bias using the likelihood ratio test (LRT) and the Wald test with ordinal data. A simulation study indicates that the scaled version of the LRT that accounts for nonnormality of the data gives untrustworthy results, whereas the unscaled LRT and the Wald test have acceptable false positive rates and perform well in terms of empirical power rate if the amount of cluster bias is large. The test for cluster bias is illustrated with data from research on teacher-student relations.     

Review of the theoretical underpinnings of loyalty programs

A review of the extant literature reveals that the theoretical underpinnings of the majority of loyalty program research rest on psychological mechanisms from three specific domains—status, habit, and relational. We propose that to understand how loyalty programs actually work, a broader, more holistic research perspective is needed to account for the simultaneous effects across these three theoretical domains as well as both cross-customer and temporal effects. The contribution of this approach is a fresh research agenda advanced in 15 research propositions.     

School bullying as a predictor of violence later in life: A systematic review and meta-analysis of prospective longitudinal studies

This paper presents results from a thorough systematic review on the efficacy of school bullying (perpetration and victimization) in predicting aggression and violence later in life. Results are based on prospective longitudinal studies. Two meta-analyses are presented examining whether: a) school bullying (perpetration and victimization) is a significant predictor of later aggression and violence, and b) whether each effect remains significant after controlling for other major childhood risk factors which were significantly related to both the predictors and the outcomes. Results are based on extensive searches of the literature. Nineteen electronic databases and 63 journals were searched from the inception of each database or journal through the end of March, 2012. Bullying perpetration at school was a significant predictor of violence (Adjusted OR = 2.04; 95% CI: 1.69–2.45) an average of six years later in life. This value of OR means that bullying perpetration increased the risk of later violence by about two-thirds. The summary effect size for bullying victimization versus violence was markedly smaller but still highly significant (Adjusted OR = 1.42; 95% CI: 1.25–1.62). This value of OR means that victimization increased the risk of later violence by about one-third. Analyses are presented of various potential moderators (such as the number of risk factors controlled for and the length of the follow-up period) in an attempt to explain the significant heterogeneity in effect sizes. Sensitivity analyses are performed on both meta-analyses, and they indicate that overall there is no evidence of publication bias. The overall findings favor the existence of a more general long-term underlying antisocial tendency rather than a more specific underlying violent tendency. Implications of our research for policy and practice are highlighted and future needs in this area of research are indicated.     

The Illusion of Exclusivity

It is widely held that when you are deliberating about whether to believe some proposition p, only considerations relevant to the truth of p can be taken into account as reasons bearing on whether to believe p and motivate you accordingly. This thesis of exclusivity has significance for debates about the nature of belief, about control of belief, and about certain forms of evidentialism. In this paper I distinguish a strong and a weak version of exclusivity. I provide reason to think that strong exclusivity is an illusion and that weak exclusivity may also be an illusion. I describe a number of cases in which exclusivity seems not to hold, and I show how an illusion of exclusivity may be generated by a rather different feature of doxastic deliberation, which I call demandingness.     

Book reviews

Freud and his Early Circle: The Struggles of Psycho-Analysis. By Vincent Brome. London: W. Heinemann Ltd. 1967. Pp. xii+275. 63s.

Fundamentals of Measurement (Second Edition). By N. M. Downie. New York: Oxford University Press. 1967. Pp. xiii + 482. 52s.

Psychological Measurement and Prediction. By Paul Horst. Belmont, California: Wadsworth. London: Prentice-Hall, 1967. Pp. vii + 455. 72s.

Attention and Performance. Edited by A. F. Sanders. Amsterdam: North-Holland Publishing Co. 1967. Pp. xii + 452. £6.

Sensory Inhibition. By Georg von Bekesy. London: Oxford University Press; Princeton University Press. 1968. Pp. x + 1265. £4 Is.

Ellements of Psychophysics, Volume I. By Gustav Fechner. Transhlated by Helmut E. Adler. New York: Holt, Rinehart & Winston Inc. 1966. Pp. xxxi + 286. 56s.

Human Information Processing. By Lawrence J. Fogel. London: Prentice-Hall. 1968. Pp. Ix + 420. 70s.

Vision and the Eye. By M. H. Pirenne. 2nd Edition. London: Chapman & Hall; New York: Barnes & Noble. 1967. Pp. xvi + 224. Hardback 50s. Paperback 25s.

The Human Operator in Complex Systems. Edited by W. T. Singleton, R. S. Easterby and D. Whitfield. London: Taylor & Francis Ltd. Pp. xx + 198. 57s. 6d. Also published as Vol. Io, No. 2, of the Journal Ergonomics.

Behavior. An Introduction to Comparative Psychology. By John B. Watson. New York: Holt, Rinehart and Winston. 1967. Pp. xxi + 439. $dollar;7.95.

The Psychology of Communication. By George A. Miller. London: Allen Lane the Penguin Press. 1968. Pp. vii + 197. 35s.     

Fitting multivariage normal finite mixtures subject to structural equation modeling

This paper is about fitting multivariate normal mixture distributions subject to structural equation modeling. The general model comprises common factor and structural regression models. The introduction of covariance and mean structure models reduces the number of parameters to be estimated in fitting the mixture and enables one to investigate a variety of substantive hypotheses concerning the differences between the components in the mixture. Within the general model, individual parameters can be subjected to equality, nonlinear and simple bounds constraints. Confidence intervals are based on the inverse of the Hessian and on the likelihood profile. Several illustrations are given and results of a simulation study concerning the confidence intervals are reported.     

A framework for ML estimation of parameters of (mixtures of) common reaction time distributions given optional truncation or censoring

We present a framework for distributional reaction time (RT) analysis, based on maximum likelihood (ML) estimation. Given certain information relating to chosen distribution functions, one can estimate the parameters of these distributions and of finite mixtures of these distributions. In addition, left and/or right censoring or truncation may be imposed. Censoring and truncation are useful methods by which to accommodate outlying observations, which are a pervasive problem in RT research. We consider five RT distributions: the Weibull, the ex-Gaussian, the gamma, the log-normal, and the Wald. We employ quasi-Newton optimization to obtain ML estimates. Multicase distributional analyses can be carried out, which enable one to conduct detailed (across or within subjects) comparisons of RT data by means of loglikelihood difference tests. Parameters may be freely estimated, estimated subject to boundary constraints, constrained to be equal (within or over cases), or fixed. To demonstrate the feasibility of ML estimation and to illustrate some of the possibilities offered by the present approach, we present three small simulation studies. In addition, we present three illustrative analyses of real data.     

Effectiveness of <Reasoning and rehabilitation> in reducing reoffending

This article aims to review the effectiveness of the program in reducing recidivism. Nineteen evaluations (involving 32 separate comparisons) were located in which experimental and control groups were compared. A meta-analysis showed that, overall, there was a significant 14% decrease in recidivism for program participants compared to controls. This program was effective in Canada and the United Kingdom but not in the United States. It was effective in community and institutional settings, whether or not it was given on a voluntary basis, and for low-risk and high-risk offenders. However, the program was more effective in older studies than in more recent evaluations.