Robert N. Sykes

Some methodological problems with feeling of knowing studies

Abstract This paper discusses two methodological problems in Feeling of Knowing (FOK) studies which emerged from a further analysis of the data of Gruneberg and Monks (1974). The first, that of greater time being spent in searching for higher FOK rated than for lower FOK rated items, is examined in an experiment which controls for differences in time of search. Results indicate that the greater search time spent on high FOK rated items in the experiment of Gruneberg and Monks could not account for differences in recall probability following first letter cueing. The second methodological problem emerging from a fuller analysis of the Gruneberg and Monks study is that the nature of the material can affect the relationship between rated feeling of knowing and subsequent cued recall.

Facilitation of receptive and productive foreign vocabulary learning using the keyword method: The role of image quality

Ellis and Beaton (1993a) reported that the keyword method of learning enhanced memory of foreign vocabulary items when receptive learning was measured. However, for productive learning, rote repetition was superior to the keyword method. The first two experiments reported here show that, in comparison with rote repetition, both receptive and productive learning can be enhanced by the keyword method, provided that the quality of the keyword images is adequate. In a third experiment using a subset of words from Ellis and Beaton (1993a), the finding they reported, that for productive learning rote repetition was superior to the keyword method, was reversed. The quality of keyword images will vary from study to study and any generalisation regarding the efficacy of the keyword method must take this into account.

The generalisability of confidence-accuracy studies in eyewitnessing

This paper discusses the evidence on the confidence-accuracy relationship in eyewitness research. It is pointed out that the conclusion often drawn on the basis of such research, that there is little or no relationship between eyewitness confidence and accuracy, is an unwarranted generalisation based on the use of experimental paradigms that are limited in terms of their generalisability to courtroom situations. In particular, almost all studies involve between-subject rather than within-subject designs, thereby limiting the generalisability of findings. A within-subjects analysis examines whether, within an individual, more confident responses are associated with greater accuracy than are less confident responses. A between-subjects analysis examines whether a more confident individual is likely to be more accurate than a less confident individual. A further limitation on the generalisability to real life situations of studies conducted to date is that experiments must involve making errors in identification in order to allow correlational analysis to take place. This means that findings cannot be generalised to those real life situations where all subjects are likely to be completely accurate and confident.

Acoustic confusion in long term memory.

Abstract Adams (1967) and Baddeley (1966a, b) amongst others, have recently claimed that short and long term memory can be distinguished on the basis of differing interference effects between the two systems, indicating that short term memory uses an auditory coding system, unlike long term memory. Specifically, it is claimed that short term memory is subject to acoustic confusion only and long term memory to semantic confusion only, despite evidence put forward by Woodworth (1938) showing acoustic confusion in long term memory. The present paper reports an experiment in which a presentation list was followed, after a period of 20–25 min., by a test list containing words acoustically similar and words acoustically dissimilar to words in the presentation list. Subjects were asked to rate, on a six point scale, whether they thought the test list word was old or new. The finding that the acoustically similar words were rated significantly higher than the non-related words, indicates acoustic confusion in LTM. In addition, the paper considers the theoretical problems involved in holding short term memory to be an auditory system.

'Coding' in Studies of Acoustic and Semantic Interference and Confusion

One of the main lines of evidence claimed to support a distinction between shortand long-term memory systems comes from the studies of immediate memory showing that acoustically related sequences are less well recalled than are semantically related sequences or nonrelated sequences (Baddeley, 1966a). From this evidence it is concluded that semantic coding does not occur in short-term memory. It was also concluded by Baddeley (1966b), on the basis of experiments on the effects of acoustic and semantic similarity of items on long-term retention, that acoustic coding is confined to short-term memory. Two questions arise from this type of experiment. What is meant by coding? And what conclusions about coding can be drawn from experiments on interference and confusion effects?