Measuring components of the memory of order

Abstract

Abstract An experienced event has many distinctively different qualitative attributes. Besides the informational content, such as which word was presented on an experimental list at a specific temporal instant, there is also information about the ordinal properties of the event relative to other events. A multinomial processing tree (MPT) model is provided for measuring four different states for the memory of the order of any arbitrary set of three events. This model, which is called the Event Order (EO) model, is linked to a specific experimental protocol for testing ordinal memory. Importantly the EO model requires neither the stipulation of a specific format for the memory encoding nor the specification of a model of order learning. The EO model is proved to be identifiable. Also the statistical estimation of the parameters of the EO model is developed using either the population parameter mapping (PPM) procedure or the maximum likelihood estimation (MLE) method, and these procedures are compared by means of Monte Carlo studies. Furthermore three experiments are described that employ the experimental protocols that enable the estimates of the parameters of the EO model. The experimental evidence shows that item knowledge is quickly acquired; whereas the corresponding acquisition of ordinal information requires many more training trials. The EO model thus provides a new way to measure ordinal information.