Recency, immediate recognition memory, and reaction time

Four classes of possible mechanisms for short-term item recognition are distinguished: (I) pure list-search, (II) direct-access activation (or trace strength) discrimination, (III) mixtures (of I and II), and (IV) response-association. Manipulations of recency, particularly of negative probe items, provide critical tests between them. Two experiments are reported using Sternberg' s 1966 varied-set reaction time paradigm, coupled with procedures intended to minimize rehearsal and control the recency of probes and memory set items. RT and error rate were greater for negative probe items that had recently been presented than for items less recently presented, and this effect increased with positive set size. In contrast, positive RT was, except for the initial item, a decreasing function of recency (= serial position), and there was no additional effect of set size per se. A brief filled delay between list and probe increased positive RT but not the slope of the set size function. These and other findings appear to reject models of Classes I and IV and, while implying some direct discrimination of an item's recency, require modification of the models of Classes II and III. The implications are discussed with respect to the relation between the two versions of Sternberg's paradigm and also in connection with facilitatory “priming” effects in memory.     

Priorities in item recognition

The Sternberg paradigm was used to examine item recognition in the experiments reported here. Functions relating reaction time to positive set size and relating reaction time to the serial position of positive targets were discussed within the context of Sternberg’s (1969) recognition model. The experiments were designed to test (a) the hypothesis that certain members of the positive set receive preferential processing and (b) the hypothesis that some members of the negative set are compared to the target in memory. The results of both experiments supported the first hypothesis, and the results of Experiment I supported the second hypothesis. Models generally consistent with these results were discussed.     

Determining subprocesses of visual feature search with reaction time models

After the classic serial/parallel dichotomy of visual search mechanisms has been increasingly doubted, we investigated what search mechanisms are used between the two poles termed "pop-out" and "strictly serial search" in an overt feature search paradigm. Since reaction time slopes do not contain sufficient information for this purpose, we developed a novel technique for analyzing reaction times. Individual reaction times are modeled as sums of the durations of successive search steps. Model parameters are task characteristics (similarity, number and arrangement of target and distractors) and processing characteristics of the participant (e.g., attention dwell and shift durations). In Experiment 1, several model variants were fitted numerically to empirical reaction times. The best fitting model suggested that more than one item can be processed in a single fixation, movement of attention is abrupt and not continuous, and even in pop out search, attention is often explicitly moved to the target. In Experiment 2, we measured the central model parameter, the so-called range of attention, more directly and thereby validated the model. The model provides an explanation for the strong variation in the slope of reaction time functions, which is not based on an explicit distinction between parallel and serial search processes.     

Effects of serial position and set size in auditory recognition memory

Auditorially presented lists of from one to six digits were memorized by Ss with instructions to scan the memory set to determine the presen’ce or absence of a single auditorially presented probe digit. Results showed cooccurrence of parallel linear increases in decision time with list length for positive and negative responses and effects of serial position of probes in the memory lists. A comparison with two previous studies using visual presentation of materials suggests faster encoding of auditorially presented probes but similarity in processing time in memory once probes have been encoded.     

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.     

Interference between maintenance and processing in working memory: the effect of item-distractor similarity in complex span

Four experiments examined the effect of phonological similarity between items and distractors on complex span performance. Item-distractor similarity benefited serial recall when distractors followed the items they were similar to, but not when distractors preceded the items they were similar to. These findings are predicted by C-SOB (contextual serial order in a box), a computational model of complex span. The model assumes that distractors are involuntarily encoded into memory, being associated to the preceding item's list position. Distractors interfere with items by superposition of distributed representations that are associated to the same position. Superposition distorts item memory; this distortion is less severe when the distractor is similar to the item. Further support for the assumption that distractors are encoded at the position of the preceding item comes from the finding that intrusions of distractors at recall tended to come from the position of the target item. In addition, intruding distractors tend to replace items to which they are similar, showing that lack of distinctiveness also contributes to interference.     

Why does the phonological similarity effect reverse with nonwords?

The detrimental phonological similarity effect (PSE), a robust finding in serial recall of words, sometimes reverses with nonwords. The current study tested the hypothesis that nonwords benefit from phonological similarity because they are harder to retrieve. In two experiments serial recall and serial reconstruction of visually presented words and nonwords were compared. Phonological similarity is known to have a positive effect on item memory and a negative effect on position accuracy in serial recall, and the demands on item retrieval were greatly reduced in the latter task. PSE occurred for words in both tasks and was reversed for nonwords in serial recall, but not in serial reconstruction—a new finding in the literature. The following conclusions can be made: (1) the detrimental PSE on order retrieval occurs irrespective of lexicality, in accordance with prominent short-term memory models; and (2), the positive PSE on item retrieval is crucially affected by lexicality, a finding less well explained by the existing models.     

A parallel-processing model for scanning

A parallel-processing model for scanning is proposed wherein all items in the display set are processed together but with different time constants. The time constant for each item depends upon its serial position and is an additive combination of primacy and recency effects. The primacy effect is linear, and the recency effect is logarithmic. Scanning can either be exhaustive or self-terminating, and “multiple looks” may occur with negative probes. It is shown that this model predicts a linear relationship between reaction time and set size, parallel or 2∶1 slope ratios depending upon the proportion of multiple looks, and serial position effects if the scanning is self-terminating.     

Scanning temporally structured lists: Evidence for dual retrieval processes

Ss were presented with lists of memory items containing one temporal pause. After each list, Ss were asked to classify a test item as positive or negative based on the memory information. In Experiments I and II, the test item was to be classified as positive if it occurred in the memory set at all, and mean reaction times for the decision were longer than for a no-pause control. In Experiment III, a positive response was required provided that the test item was in a relevant portion of the list as defined by a postlist cue. Reaction times were longer and showed a somewhat different pattern than for a no-pause control. An interpretation was offered based on the hypothetical existence of two simultaneously executed serial scans, one "exhaustive" and the other "self-terminating," as defined by S. Sternberg.     

Modulating the phonological similarity effect: The contribution of interlist similarity and lexicality

The classical phonological similarity effect (PSE) was studied with words and nonwords in two immediate serial recall (ISR) tasks. The relative contributions of intralist and interlist interference were compared, and differential effects on item and order memory were observed. PSE occurred with words and was reversed with nonwords. In addition, PSE was modulated by interlist similarity, which enhanced recall of rhyme items and impaired recall of distinct items. Finally, interlist similarity reduced item recall of words, whereas it improved serial recall of nonwords. The latter finding rules out the hypothesis that the reverse PSE for nonwords is due to interlist interference. It is concluded that two opposing effects of phonological intralist similarity cause the interaction between PSE and lexicality in ISR. With words, the positive effect on item recall is usually masked by a much more disruptive effect on position accuracy. With nonwords, however, the positive effect often masks the negative one. These findings are discussed in relation to current models of verbal short-term memory.     

Beyond similarity: Masking of the target is sufficient to cause the attentional blink

When subjects are asked to identify a letter target embedded in a rapid serial visual presentation stream, the detection of a subsequent letter probe is briefly impaired. This transient deficit in probe detection, termed the “attentional blink,” depends on the type of item that immediately follows the letter target (Raymond, Shapiro, & Arnell, 1995). Two models have been proposed to account for this effect. The interference model of the attentional blink predicts that visual similarity between the probe and item immediately following the target (+1 item) causes the attentional blink, whereas the two-stage model is based on the notion that increased time needed to process the target letter causes the attentional blink. In order to test between these two possibilities, the masking properties of the +1 item and its similarity to the probe were varied. We found the attentional blink when the +1 item acted as a mask of the target, even though the +1 item and the probe were visually dissimilar. This pattern of results supports the two-stage model of the attentional blink.     

On the interaction between memory scanning and response set

In the first of two item recognition experiments, a study by Egeth, Marcus, and Bevan (1972) was duplicated, but using a within-Ss design. Consistent with their findings, analogous differences obtained between mean RTs to positive set sizes of one and three were found to be significantly greater when Ss pressed two different keys to indicate whether or not a test letter was a member of the positive set (two-response condition) than when Ss pressed a single key to indicate only that a test letter was a positive set member (one-response condition). The number of set sizes used and results obtained for the two-response condition, which were inconsistent with characterization of the search process as either serial and exhaustive or serial and self-terminating, both limited and complicated interpretation of the obtained interaction. Further, these results suggested that the procedures used differed in important ways from those known to consistently produce linear and parallel item recognition functions for both positive and negative trials for the two-response condition. In Experiment II, those item recognition procedures which are known to allow serial exhaustive search to occur were used, and data were collected for four positive set sizes. Reaction time was a linear function of set size for both response conditions, the slopes of the functions did not differ significantly, and absolute slope values were consistent with previously reported findings for the two-response condition and letter stimuli. It is concluded that response requirements sometimes have effects on memory scanning (the slope of the item recognition function). Such an effect was observed when the experimental conditions were such as to allow neither serial and exhaustive nor serial and self-terminating memory search to occur. But, response requirements do not always have effects on memory scanning, and this was found to be the case when the experimental conditions were such as to allow serial and exhaustive memory search to occur and it was shown to occur.     

Investigating mental representation of order with a speeded probed recall task

We compare three models of representation of item order in a verbal STM task: item-item associations, item-position associations, and primacy gradient. A speeded probed recall task is used, in which a list of words is presented, immediately followed by a probe; participants must report as fast as possible the word that was in the probed position. In the number probe condition, a digit is presented and one must say the word in that position. In the word probe condition, the probe is an item of the list and participants must say the immediately following item. Response times (RTs) are analyzed according to probe type and position. The three models imply different predictions about RTs as a function of serial order in the two conditions. Our results suggest a serial, self-terminating search from the beginning of the list to the target position, except for the final position, which is directly accessible. The item-item and item-position association models are ruled out; the primacy gradient model accounts satisfactorily for our results, except for the finding of a larger recency effect with a number probe. Alternative interpretations are also discussed.     

Investigating a memory-based account of negative priming: support for selection-feature mismatch

Using typical and modified negative priming tasks, the selection-feature mismatch account of negative priming was tested. In the modified task, participants performed selections on the basis of a semantic feature (e.g., referent size). This procedure has been shown to enhance negative priming (P. A. MacDonald, S. Joordens, & K. N. Seergobin, 1999). Across 3 experiments, negative priming occurred only when the repeated item mismatched in terms of the feature used as the basis for selections. When the repeated item was congruent on the selection feature across the prime and probe displays, positive priming arose. This pattern of results appeared in both the ignored- and the attended-repetition conditions. Negative priming does not result from previously ignoring an item. These findings strongly support the selection-feature mismatch account of negative priming and refute both the distractor inhibition and the episodic-retrieval explanations.     

Evidence for direct visual access to letter identities

Subjects were presented with two letters from the set BCDFGHLNPORTVZ, one in upper case and one in lower case. Subjects were required to decide, as quickly as possible, whether the two letters on each trial were the same (e.g., Bb) or different (e.g., Bd). Because letters were always physically different, subjects presumably must name the letters in order to respond. Reaction times were predicted by the visual similarity of the letters, and not by their phonemic distinctive feature similarity. Consistent with previous work showing that a word's meaning can be assessed directly from visual information, the findings suggest that visual analysis of a capital or small letter can result in knowing the latter name without further linguistic processing. These results were interpreted as an example of automatic detection as described by Schneider and Shiffrin (1977).     

A re-evaluation of the phonological similarity effect in adults' short-term memory of words and nonwords

The phonological similarity effect (PSE) was studied in two tasks of serial recall, in one task of serial recognition and one item identification task. PSE occurred only in the former three tasks involving memory of order when study items were words and nonwords with an associative connectedness to long-term memory. Nonwords that, according to a reaction time assessment of associative value, were less well connected to long-term memory mechanisms, were not sensitive to phonological similarity. These results are discussed in relation to contemporary models of short-term memory that explain the PSE as a result of confusions of items that are similarly encoded in a phonological layer. This layer is identified as a higher-level phonological space that is accessed by words and nonwords of high associative value and not by nonwords of low associative value.     

Generative processes in character classification: Evidence for a probe encoding set

The role of rehearsal in a varied set memory scanning task was investigated by asking the subjects to rehearse subvocally the one to four target characters cyclically at a self-paced rate until the probe was presented. After making a manual positive or negative response to the probe, the subjects reported the last item rehearsed before the probe was presented. The results indicate that, when the last rehearsed item matched the probe, RTs were significantly faster than when it differed. Mean RTs over target set size were generally well fit by linearly increasing RT functions, with comparable slope values for negative responses and positive responses when the last rehearsal was the same as or different from the probe. The data suggest that rehearsal may reduce the duration of the probe encoding stage through some mechanism of pathway activation by providing the subject with a generated representation of what may appear next as a probe.     

Overwriting of phonemic features in serial recall

We tested two explanations of the phonological similarity effect in verbal short-term memory: The confusion hypothesis assumes that serial positions of similar items are confused. The overwriting hypothesis states that similar items share feature representations, which are overwritten. Participants memorised a phonologically dissimilar list of CVC-trigrams (Experiment 1) or words (Experiment 2 and 3) for serial recall. In the retention interval they read aloud other items. The material of the distractor task jointly overlapped one item of the memory list. The recall of this item was impaired, and the effect was not based on intrusions from the distractor task alone. The results provide evidence for feature overwriting as one potential mechanism contributing to the phonological similarity effect.     

Encoding location and serial order in auditory working memory: evidence for separable processes

In this study, we investigated the interactions between temporal and spatial information in auditory working memory. In two experiments, participants were presented with sequences of sounds originating from different locations in space and were then asked to recall either their position or their serial order. In Experiment 1, attention during encoding was manipulated by contrasting 'pure' blocks (i.e., location-only or serial-order-only trials) to 'mixed' blocks (i.e., different percentages of spatial and serial-order trials). In Experiment 2, 'pure' blocks were contrasted to blocks in which spatial and serial-order trials were intermixed with a third task requiring a semantic categorization of sounds. Results from both experiments showed that, whereas serial-order recall is linearly affected by the simultaneous encoding of a concurrent feature, the recall of position is mostly unaffected by concurrent feature encoding. Contrastingly, overall performance level was lower for spatial recall than serial recall. We concluded that serial order and location of items appear to be independently encoded in auditory working memory. Serial order is easier to recall, but strongly affected by the processing of concurrent item dimensions, while item location is more difficult to recall, but relatively automatic, as shown by its strong resistance to interfering dimensions in encoding.