Ray Pike

Global matching: A comparison of the SAM, Minerva II, Matrix, and TODAM models.

The pair recognition paradigm was chosen as a means to evaluate and compare different ways to implement a global matching process in which the matches with all list items are combined into a single value. Simplified versions of the SAM, Minerva II, Matrix, and TODAM models are shown to be specific instances of a more general model and are shown to share a parameter-free prediction. This prediction is then shown to be unaffected by the inclusion of additional processes into the models such as variable encoding, cue weights, forgetting, single item matches, context, and background memories. Furthermore, it is only slightly affected by the formation of interpair associations. A test of the prediction using word pairs was inconclusive due either to the use of a recall strategy or to a low level of interitem similarity. Ways to differentiate between the different models and to test the global matching assumption rigorously are discussed.

Associative interference effects in recognition and recall

The results of three studies testing whether associative interference occurs in recognition as it does in recall are reported.

A multiple-observations model for response latency and the latencies of correct and incorrect responses in recognition memory.

A model for response latency in recognition memory is described which is a strength model incorporating the notion of multiple observations and with the additional assumptions that the variance of the strength distributions increase with set size and that the observer attempts to keep his error rate at a constant level over set size. It is shown that the model can, without recourse to particular parameter values, predict a near linear RT set-size function and, since it is a (TSD) model in its decision aspects, can account for errors and hence error latencies in the recognition task. After the model is described, two experiments are performed which test the prediction that correct mean latency is generally shorter than incorrect mean latency. The prediction is confirmed and this feature is discussed in general, the model being compared with that of Juola, Fischler, Wood, and Atkinson (1971) in this respect. Some possible modifications to the latter model are also considered.

Response latencies in the yes/no detection task: An assessment of two basic models

Two general forms of model for response latency in the yes no signal detection task are described and their predictions examined in the light of experimental data. One model supposes a latency function along the detection axis and the other supposes that a count is made of multiple observation, Experiment I looks at the auditory detection situation with short observation interval, but the results are generally equixocal, although the latency function model gives a simpler interpretation Other results are then discussed, particularly those of Gescheider et al (1969). and it is concluded that an attempt must be made to replicate the results of Carterette et al (1965) for the extended observation interval detection experiment. This is achieved in Experiment II. and the obtained order of mean latencies is then examined in terms of the models The counting model appears by far the more suitable for this situation.

Using multilist designs to test for contextual reinstatement effects in recognition

Gillund and Shiffrin (1984) proposed an explanation for the lack of a contextual matching advantage in recognition. We found that although their theory can predict no effect in a single-list design, it generally predicts an effect in a multilist design. An experiment to test this prediction was inconclusive, because the contextual manipulation used (study room) was not shown to have reliable effects.

What is the signal in signal detection

This study concerns the nature of the stimulus represented along the decision axis in the yes/no auditory detection task. Two contrasting interpretations, absolute and difference representation of the stimulus, are tested by raising the carrier tone embedded within the background “noise” to the level of the signal, on occasional “catch” trials. Results indicate that difference detection may be the preferred mode of operation when a carrier tone is present and the task is a difficult one. Implications for the TSD model are discussed in terms of the relative efficiency of the two detection mechanisms.

Further analysis of the Crist data on matching performance and the similarity structure of the stimulus set.

We argue that Crists (1981) analysis of his data on letter matching is incomplete because he did not sufficiently distinguish overall context similarity from the similarity of the other letter present on a trial. That is, the separation of the effects of context similarity and visual similarity is required for a complete analysis. Accordingly, for different trials, modified indices of context similarity are derived and linear regression analyses made to determine the separate effects of context and visual similarity. The results of these analyses show that on different trials, visual similarity is much more predictive of latency than context similarity and that subjects must therefore be making visual similarity judgments, that considerably affect latency variation on these trials. Consequently, we argue that visual similarity judgment may be likewise affecting latency on same trials. However, some simple correlations indicate that context may be more predictive of latency on physically different same trials. We concluded that what can be said about context depends on the type of trial and that the comparative strength of name processing should also be assessed by the method described here.

Order and serial position effects in response time with multiple-item probe recognition

We examined the order effect in item-recognition response time, that is, differences in response time for multiple-item probes containing items in the same or in the reverse order as those in the memory set. Experiment 1 used the response condition in which only one item must be positive for a positive response, Experiment 2 used homogeneous probes in which all the items are either positive or negative, and Experiment 3 used the condition in which all the items must be positive. Of particular interest were the serial position variations in order effects for probes containing items that were adjacent in the memory set. We previously found that such effects are an indication of subjective grouping of the memory set and the matching of the probe with these subgroups. The order effect in the one-positive condition was only weak in most cases, but it was strong with homogeneous probes when the memory set was objectively grouped or was ungrouped but with a constant set size. There were also strong order effects in the all-positive condition for probes with items that were nonadjacent in the memory set. Our results are interpreted in terms of a parallel match process based on a distribution over position of items in subjective or objective groups. We account for the origin of the distribution-over-position process in terms of multiple representations of the grouped memory sets. The model assumes that each subgroup is represented in memory several, and perhaps very many, times and that considerable error in item positioning can occur over the multiple representations of any group.

The components of latency of response in two models for auditory detection with deadlines

Two models are described for the case of auditory detection with response time deadlines. These are the “timing” model, modified from Luce and Green (1972), and the “interval of uncertainty” (IU) model, modified from the “counting” model of McGill (1967). The two models are expressed in the form of additive component latencies, including a component which varies inversely with probability of response and a component for a longer response on the “nondeadline” response condition. An experiment is described, incorporating deadlines on both s and n trials separately, and the additive component latencies of the models are calculated. The “interval of uncertainty” model is seen to represent the data better than the “timing” model, even though different versions of the latter model are considered. Parameters are then fitted to give a representation of the IU model. The data from one subject of another study not employing deadlines can also be fitted to the model, and an interesting problem then arises concerning the exact locus of the speed-accuracy tradeoff. It is concluded that the IU model needs extending if it is to describe the data without the use of a component of latency which varies inversely with response probability.

Criteria placement in detection-implications for kogan and wallach's theory of risk-taking

Abstract Consideration of Kogan and Wallachs (1964) theory of risk-taking leads to the prediction that Ss high in Test Anxiety and Defensiveness should be consistently risky or consistently conservative in their criteria placement in detection. Ss low on these measures should tend to behave optimally in these tasks. Results indicate that while low TA/Def Ss are responsive to task demands, in as much as they change criteria in the correct direction, they do not tend to perform optimally. High TA/Def Ss only behave conservatively in the detection situation. The generality of Kogan and Wallachs model is questioned and some logical difficulties discussed.