Thorsten Meiser

Measuring source memory

Abstract. The investigation of source monitoring (SM) as a special faculty of episodic memory has gained much attention in recent years. However, several measures of source memory have been used in research practice that show empirical and theoretical shortcomings: First, they often confound various cognitive processes like source memory, item memory and response bias, and second, they do not do justice to the multitude of processes involved in SM according to the framework of Johnson, Hashtroudi, and Lindsay (1993). We therefore review model-based measurement approaches, focusing on multinomial models, and we distinguish between theorizing about source memory and the pragmatics of source memory measurement as two partly separate goals of research. Whereas signal detection models seem to be more adequate theories of the underlying source monitoring process, multinomial models have some pragmatic advantages that nevertheless recommend them as viable measurement tools.

Binding of multidimensional context information as a distinctive characteristic of 'remember' judgments

This research investigated the cognitive processes underlying remember-know judgments in terms of contextual binding in multidimensional source memory. Stochastic dependence between the retrieval of different context attributes, which formed the empirical criterion of binding, was observed for remembered items but not for known items. Experiment 1 showed that the qualitative difference in the stochastic relation holds even if quantitative source-memory performance is equated for items with remember and know judgments. Experiment 2 generalized the findings to context information from different modalities, and Experiment 3 ruled out a spurious stochastic dependence due to interindividual differences. Supporting recent dual-process models of remember-know judgments, the findings show that remember and know judgments differ with respect to binding processes that correspond to episodic recollection.

A Source-Monitoring Analysis of Illusory Correlations

Source monitoring refers to the discrimination of the origin of information. The source-monitoring methodology, applied to illusory correlations in the formation of stereotypes, allows one to disentangle memory for behaviors from memory for the behaviors’ group origin and from response bias. In three studies, illusory correlations are found, and they are shown to reflect differential response bias rather than differential item or group memory. In addition, illusory correlations are found only along an evaluative dimension, not for a gender classification of group members. The results challenge so-called cognitive accounts of illusory correlations, such as the account by distinctiveness, whereas they can be reconciled with an account in terms of evaluative differentiation.

Cognitive processes in stereotype formation: the role of correct contingency learning for biased group judgments.

Three studies investigated contingency learning and stereotype formation in a scenario about group membership and behavior with a confounding context factor. The studies tested predictions from theoretical accounts of biased group judgments in terms of simplistic reasoning, parallel distributed memory, and pseudocontingencies. Study 1 revealed a positive correlation between erroneous stereotype formation and learning of the true contingencies with the confounding factor. Study 2 showed that a focus manipulation during encoding moderated the correlation between stereotype formation and contingency learning but not the strength of the erroneous stereotype. Study 3 used a quasiexperimental comparison between participants with biased versus unbiased group judgments and extended the findings of a positive relation between stereotype formation and contingency learning. The results support an explanation of biased group judgments by pseudocontingencies; that is, unwarranted inferences from accurately perceived bivariate correlations in complex environments.

Metacognitive inferences in source memory judgements: The role of perceived differences in item recognition

This research investigated the hypothesis that metacognitive inferences in source memory judgements are based on the recognition or nonrecognition of an event together with perceived or expected differences in the recognizability of events from different sources. The hypothesis was tested with a multinomial source-monitoring model that allowed separation of source-guessing tendencies for recognized and unrecognized items. Experiments 1A and 1B manipulated the number of item presentations as relevant source information and revealed differential guessing tendencies for recognized and unrecognized items, with a bias to attribute unrecognized items to the source associated with poor item recognition. Experiments 2A and 2B replicated the findings with a manipulation of presentation time and extended the analysis to subjective differences in item recognition. Experiments 3A and 3B used more natural source information by varying type of acoustic signal and demonstrated that subjective theories about differences in item recognition are sufficient to elicit differential source-guessing biases for recognized and unrecognized items. Together the findings provide new insights into the cognitive processes underlying source memory decisions, which involve episodic memory and reconstructive tendencies based on metacognitive beliefs and general world knowledge.

The role of metacognition in prospective memory: Anticipated task demands influence attention allocation strategies

The present study investigates how individuals distribute their attentional resources between a prospective memory task and an ongoing task. Therefore, metacognitive expectations about the attentional demands of the prospective-memory task were manipulated while the factual demands were held constant. In Experiments 1a and 1b, we found attentional costs from a prospective-memory task with low factual demands to be significantly reduced when information about the low to-be-expected demands were provided, while prospective-memory performance remained largely unaffected. In Experiment 2, attentional monitoring in a more demanding prospective-memory task also varied with information about the to-be-expected demands (high vs. low) and again there were no equivalent changes in prospective-memory performance. These findings suggest that attention-allocation strategies of prospective memory rely on metacognitive expectations about prospective-memory task demands. Furthermore, the results suggest that attentional monitoring is only functional for prospective memory to the extent to which anticipated task demands reflect objective task demands.

Loglinear Rasch Models for the Analysis of Stability and Change.

Loglinear unidimensional and multidimensional Rasch models are considered for the analysis of repeated observations of polytomous indicators with ordered response categories. Reparameterizations and parameter restrictions are provided which facilitate specification of a variety of hypotheses about latent processes of change. Models of purely quantitative change in latent traits are proposed as well as models including structural change. A conditional likelihood ratio test is presented for the comparison of unidimensional and multiple scales Rasch models. In the context of longitudinal research, this renders possible the statistical test of homogeneity of change against subject-specific change in latent traits. Applications to two empirical data sets illustrate the use of the models.

A Hierarchy of Multinomial Models for Multidimensional Source Monitoring

Abstract. Several models have been proposed for the measurement of cognitive processes in source monitoring. They are specified within the statistical framework of multinomial processing tree models and differ in their assumptions on the storage and retrieval of multidimensional source information. In the present article, a hierarchical relationship is demonstrated between multinomial models for crossed source information (Meiser & Broder, 2002), for partial source memory (Dodson, Holland, & Shimamura, 1998) and for several sources (Batchelder, Hu, & Riefer, 1994). The hierarchical relationship allows model comparisons and facilitates the specification of identifiability conditions. Conditions for global identifiability are discussed, and model comparisons are illustrated by reanalyses and by a new experiment on the storage and retrieval of multidimensional source information.

The Personal Structure of Personal Need for Structure A Mixture-Distribution Rasch Analysis

Abstract. This study investigated the responses of N = 1,789 participants to a set of 12 Likert-type items for the assessment of personal need for structure (PNS). Mixture-distribution Rasch models were used to analyze the homogeneity of the response format across items and the homogeneity of the item parameters and category parameters across persons. Model selection yielded a two-class rating scale model as the favorite model. This model contains the assumptions that the Likert response scale is used in a constant way for all items but that the item or category parameters differ between two latent subpopulations. The parameter estimates revealed large differences in the threshold parameters for the response categories between the two subpopulations. While the larger subpopulation showed a tendency to avoid extreme response categories, the smaller subpopulation used the whole range of the response scale. The different response styles identified by the mixture-distribution Rasch analysis were validated by ...

On the automatic nature of the task-appropriate processing effect in event-based prospective memory

This research tested competing predictions about the cognitive processes underlying the task-appropriate processing effect in prospective memory. Participants had to press a designated key whenever a word from the semantic category of animals or from the structural category of palindromes occurred in an ongoing lexical decision task. The availability of attentional resources was manipulated by varying the effort to the ongoing task in terms of speed or accuracy. In the task-appropriate semantic prospective memory task, performance was robust against the speed versus accuracy instructions. In the task-inappropriate structural prospective memory task, performance declined under speed instructions that detracted attentional resources. Accordingly, a facilitating effect of task-appropriate processing was observed under speed instructions but not accuracy instructions. The results support the notion that the task-appropriate processing effect is due to a larger contribution of automatic cue detection to prospective memory performance under task-appropriate than task-inappropriate conditions.