Volume 46, Issue 2 (2004): Contents

This issue opens with a paper by Dieter Rasch and Volker Guiard on how well parametric statistical methods hold up when their assumptions are not fully met. The question of robustness runs quietly through several of the other contributions too.

Erling B. Andersen contributes a piece on latent regression analysis built on the rating scale model, extending that framework to settings where latent traits are linked to background variables.

A second paper from Dieter Rasch, written with Klaus D. Kubinger, Jörg Schmidtke and Joachim Häusler, takes aim at a habit many readers will recognize: the scattering of asterisks across tables to flag significance levels. The authors treat this shorthand as a misuse of hypothesis testing rather than a harmless convention.

Alexander von Eye and G. Anne Bogat address a problem that sits underneath a good deal of multivariate work, namely how to actually test whether data satisfy the assumption of multivariate normality instead of taking it for granted.

Michaela M. Wagner-Menghin reports on the content validity of an objective personality test designed to measure the achievement motive, examining whether the instrument captures what it claims to.

The issue closes with Wulf-Uwe Meyer, Rainer Reisenzein and Oliver Dickhäuser on a subtler bit of social cognition: how people infer someone's ability from the blame directed at them, and how effort-oriented versus liking-oriented cognitive schemata shape that inference.