In this article, we analyze both M. J. Wainwright and E. M. Reingold's view of the process dissociation measurement models presented by A. Buchner, E. Erdfelder, and B. Vaterrodt-Plunnecke and their suggestions on that topic. This analysis reveals a number of problems in Wainwright and Reingold's approach. Some of these problems are more subtle than others, but they are nevertheless consequential. Thus, researchers working with the process dissociation procedure should be aware of these problems.

In two experiments the conditions of, and the processes leading to, gender biases in fame judgments were investigated. In Experiment 1, the gender bias was not reduced in a condition that alerted participants to the gender of the names. In Experiment 2, participants' sex-role orientation, but not their gender, was related to the gender bias. The process dissociation procedure was used in both experiments in an attempt to separate conscious and unconscious memory processes contributing to the gender bias. Using L. (...) L. Jacoby's original measurement model there appeared to be evidence for unconscious influences on the gender bias in fame judgments. Unfortunately, this evidence disappeared when a model was used that takes guessing and, hence, response biases into account, which confirms that measurement models that ignore response biases in the process dissociation procedure may lead to erroneous conclusions. (shrink)

Previous research has shown a source memory advantage for faces presented in negative contexts. As yet it remains unclear whether participants remember the specific type of context in which the faces were presented or whether they can only remember that the face was associated with negative valence. In the present study, participants saw faces together with descriptions of two different types of negative behaviour and neutral behaviour. In Experiment 1, we examined whether the participants were able to discriminate between two (...) types of other-relevant negative context information (cheating and disgusting behaviour) in a source memory test. In Experiment 2, we assessed source memory for other-relevant negative (threatening) context information (other-aggressive behaviour) and self-relevant negative context information (self-aggressive behaviour). A multinomial source memory model was used to separately assess partial source memory for the negative valence of the behaviour and specific source memory for the particular type of negative context the face was associated with. In Experiment 1, source memory was specific for the particular type of negative context presented (i.e., cheating or disgusting behaviour). Experiment 2 showed that source memory for other-relevant negative information was more specific than source memory for self-relevant information. Thus, emotional source memory may vary in specificity depending on the degree to which the negative emotional context is perceived as threatening. (shrink)

In this article, we try to clarify some of the issues raised by S. C. Draine, A. G. Greenwald, and M. R. Banaji concerning our investigation into the gender bias in fame judgments . First, we did not test the general hypothesis and did not draw the general conclusion that Drain et al. suggest we did. Second, we did not reject M. R. Banaji and A. G. Greenwald's assumptions about the familiarity of male and female names in the fame judgment (...) task, but we showed how one could have come to reject it using a widespread measurement model for the process dissociation procedure. Third, we argue that the processes which Draine et al. suggest should also be included in the measurement model we used are probably negligible, and if they are not, then the validity of the results of a number of fame judgment experiments must be called into question. In general, however, we agree with what seems to be the main message ofM. R. Banaji and A. G. Greenwald's research, namely, that social categories have to be considered whenever priming is investigated within a social domain. (shrink)