Wayne A. Wickelgren

Speed-accuracy Tradeoff and Information Processing Dynamics

Abstract For a long time, it has been known that one can tradeoff accuracy for speed in (presumably) any task. The range over which one can obtain substantial speed-accuracy tradeoff varies from 150 msec in some very simple perceptual tasks to 1,000 msec in some recognition memory tasks and presumably even longer in more complex cognitive tasks. Obtaining an entire speed-accuracy tradeoff function provides much greater knowledge concerning information processing dynamics than is obtained by a reaction- time experiment, which yields the equivalent of a single point on this function. For this and other reasons, speed-accuracy tradeoff studies are often preferable to reaction-time studies of the dynamics of perceptual, memory, and cognitive processes. Methods of obtaining speed-accuracy tradeoff functions include: instructions, payoffs, deadlines, bands, response signals (with blocked and mixed designs), and partitioning of reaction time. A combination of the mixed-design signal method supplemented by partitioning of reaction times appears to be the optimal method.

Strength models and serial position in short-term recognition memory ☆

Abstract A number of continuous strength models for memory are developed for and tested by an experimental study of recognition memory for three-digit numbers at all serial positions in lists of length two through seven. Empirical estimates of trace strength in different conditions, independent of response bias, are obtained by means of the operating characteristic. The principal theoretical findings are: (a) strength in short-term memory (STM) appears to decay exponentially with the number of subsequent items; (b) subjects report that they recognize an item if and only if strength in memory exceeds a criterion; (c) the first item of a list is remembered better than subsequent items because it receives a greater increment in strength in STM upon presentation, not because it decays more slowly in STM or because it acquires some strength in a long-term memory.

Single-trace fragility theory of memory dynamics

In single-trace fragility theory, forgetting is produced by two factors, time and interference. Memory traces are assumed to have two partially coupled dynamic properties, strength and fragility. Strength determines the probability of correct recall and recognition, while fragility determines the susceptibility of the trace to the time-decay process but not to the interference process. Consolidation is assumed to be a continual reduction in the fragility of the memory trace rather than any change in strength or availability. Decreasing fragility accounts for the continually decreasing forgetting rate, the temporal character of retrograde amnesia and recovery therefrom, and the type of internal clock necessary for nonassociative recency judgments. Data are presented to indicate that interference is independent of the interval between original and interpolated learning, that nonassociative recency discriminability approaches a limit at about 30 min, and that the decay rate of long-term retention in amnesic patients is the same as in normal Ss.

Distinctive Features and Errors in Short‐Term Memory for English Consonants

Errors in short‐term recall of 23 English consonants were tabulated and related to three distinctive‐feature systems. The consonants were always presented in initial position in a consonant‐vowel diagram, and the vowel was always /a/. Subjects were instructed to copy a list of consonants as it was being presented, followed by recall of the list. Perceptual errors were excluded from the recall‐error matrix by scoring for recall only correctly copied consonants. The data were also analyzed in such a way as to eliminate differences in response bias for different consonants. Having controlled for response bias, each feature system makes predictions about the rank order of different intrusion errors in recall. Each of the three feature systems was significantly more accurate than chance in these predictions, but the most accurate system was one developed in the present study. This system is a slightly modified version of the conventional phonetic analysis of consonants in terms of voicing, nasality, openness o...

Trace resistance and the decay of long-term memory.

Abstract A theory of storage in long-term memory is described which characterizes the trace by two properties: strength and resistance. The theory makes four assumptions: (a) The rate of decay of strength equals the force on the trace divided by the resistance. (b) Force is proportional to trace strength and to the similarity of current traces to the previously established trace. (c) Resistance increases as the square root of trace age. (d) The resistance of a trace transfers completely to subsequent increments. The theory accounts for long-term retention functions over delays from 1 min to 2 yr, long retrograde amnesia, unlearning, effects of multiple learning trials and spacing of practice, and possibly recency judgments and retention functions following relearning.

Sparing of short-term memory in an amnesic patient: Implications for strength theory of memory ☆

Abstract Short-term recognition memory for single-digit numbers, three-digit numbers, and the pitch of pure tones was studied in a subject (H.M.) who appears to possess normal short- term memory, but to be almost completely lacking in the ability to form new long-term memory traces. The primary purpose was to test the short-term memory component of a dual-trace strength theory developed for normal subjects. To a first approximation, H.M.s trace-strength decay curves are exponential, as postulated by strength theory for short-term memory. Furthermore, the rate of decay of this trace for H.M. is well within the range of normal subjects. These findings agree with previous findings on H.M. in support of a dual-trace theory of memory. More particularly, the present findings support the strength theory of memory proposed by WICKELGREN and NORMAN [15].

Associative strength theory of recognition memory for pitch

Abstract Recognition memory for pitch was studied by means of a delayed comparison task, with the standard (S) tone and the comparison (C) tone separated by a variable delay interval (0–180 sec). Evidence is presented for the existence of an unsigned familiarity or similarity dimension, in addition to the signed pitch-difference dimension. Subjects relied on familiarity exclusively in the same-different judgment and used both dimensions in the higher-same-lower judgment. There appear to be two memory traces, short-term and intermediate-term, which are decaying exponentially to zero at very different rates. The decay of the short-term trace appears to be an essentially passive, temporal decay process, rather than an interference process, since the frequency-similarity and intensity of the interference tone have no effect on the rate of decay. The decay of the memory trace appears to have the same form and rate regardless of the initial level of acquisition and regardless of the frequency difference between the S and C tones.

Unidimensional strength theory and component analysis of noise in absolute and comparative judgments

Abstract Using the decision rules and normal distribution assumptions of signal-detection theory as a base, a general strength theory of unidimensional absolute and comparative judgments is described in detail. The componenss of variance in both absolute and comparative judgments are considered, with particular emphasis on criterion variance in an absolute-judgment task and its relation to criterion variance in a comparative-judgment task. Some difficulties are noted in predicting comparative-judgment (forced-choice) probabilities from absolute-judgment (“yes-no”) probabilities. The principal difficulties are concerned with the relative magnitudes of criterion variance in the two tasks, the correlation of distributions, and attention. The question of the equality of variances for different criteria (e.g., yes-no vs confidence criteria) is considered, and two methods are suggested for answering the question (one of which is a new type of operating characteristic). The notion of a random variable being a function of a real variable or being a function of another random variable is used to analyze the effects of noise in an independent variable on the distribution of a dependent random variable.

Acoustic similarity and retroactive interference in short-term memory

Summary Short-term memory for a list of four letters, followed by a list of eight letters that the S s copied as they were presented, followed by immediate recall of the original four-letter list, was shown to be a function of the acoustic similarity of the intervening list to the original list. An interfering list whose letters have similar pronunciation to the letters in the original list produces greater RI than an interfering list whose letters have a very different pronunciation from the letters in the original list. An interfering list composed of items identical to items in the original list, but in a different order, tends to produce less RI in the recall of items and more RI in the recall of the correct position of these items than an interfering list composed of similar items. These findings for STM are completely consistent with analogous studies of RI as a function of similarity in LTM.

Strength theory of decision rules and latency in retrieval from short-term memory.

A simple strength theory is studied for retrieval in recognition, multiple-choice and recall tests of verbal short-term memory. The primary assumptions of the theory are: (a) that the same memory system is used, regardless of the manner of testing retention, (b) that in recognition, subjects respond “yes” if the strength of the test item exceeds a criterion (criterion decision rule), and (c) that in multiple-choice or recall, subjects choose the response with the greatest strength (maximum decision rule). The results support these basic retrieval assumptions, but suggest that the decay and noise in retrieval may not always be constant for different methods of testing retrieval. Response latencies were studied as a subsidiary measure of trace strength.