Unlearning in recognition memory

Garskof, Bertram E.; Sandak, Joel M.

doi:10.3758/bf03342861

Cited by 9 publications

(2 citation statements)

References 1 publication

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The present results differ from those of Garskof & Sandak (1964) and Postman (1962) who reported poorer associative matching (a kind of recognition test) of A-B following A-C learning than following a rest interval. The experiments differ in at least the following ways: (a) here A-B, A-C was compared to an A-B, C-D control rather than to an A-B-rest control; (b) our Ss merely studies each pair once, whereas their Ss had many overt anticipation trials with lists of such pairs; and (c) their associative matching test elicits complex list strategies for a set of items of varying strengths, whereas our test obtained a recognition rating for each pair separately, half tested positively and half negatively.…”

Section: Discussioncontrasting

confidence: 57%

Absence of within-list PI and RI in short-term recognition memory

Bower

Bostrom

1968

Psychon Sci

View full text Add to dashboard Cite

The question of interest here is whether unlearning of specific competing S-R associations can be demonstrated in short-term memory by recognition tests. Whether specific interference is expected in recognition tests depends on our conception of how pair recognition occurs as well as whether we assume that learning A-C weakens the A-B association. Plausible alternatives for recognition of A-B following A-C learning are: (1) S searches through memory until he finds either associate to A, giving the recognition R only if the first one he finds matches the test pair; (2) S exhaustively searches for all Rs associated to A, and gives the recognition R only if the test R is the "strongest" of the Rs retrieved for A; or (3) S retrieves the memory trace of the test SR pair and gives the recognition R only if its strength exceeds a criterion. If A-C causes unlearning of an initial A-B association, then either of these rules predicts Rl effects. If A-C and A-B are stored independently, then RI and PI effects are still expected by Rules 1 and 2, but not by 3.Rule 3 assumes that recognition of a given A-B pair depends only on its absolute strength and is independent of other associations to A. The following experiment was conducted to determine whether RI and PI effects would appear in short-term recognition memory. MethodThe experiment consisted of a study-then-test block on each of 15 different lists of S-R pairs. The stimuli were consonant bigrams of minimal intralist similarity, different in every list, and the responses in each list were all the digits 1-9. Each study list consisted of 17 pairs, divided into the first three "primacy buffer" items (never tested) followed by two sUblists each of seven items as illustrated in Table 1. Three "interference" items occurred in Sublist 1, were repeated Psychon. Sci., 1968, Vol. 10 (6) GORDON H. BOWER AND ALAN BOSTROM STANFORD UNIVERSITYin Sublist 2 with a different response, and were then tested for recognition: for one item (RI+) the firstpaired response was tested positively, for another (PI+) the second-paired response was tested positively, and for the third (1-) the test response was incorrect. The "negative" recognition tests with incorrect pairs were needed to assess false alarm rates. Two RI control items appeared in Sublist 1 but not in Sublist 2, and two PI control items appeared in Sublist 2 but not in 1. Two Repetition items were included withthe same pairings in both sublists; this was to gauge sensitivity since any theory predicts benefits for such pairs. Items within each sublist were randomly ordered except that at least three items intervened between occurrences of the same stimulus. Following the study block, the nine tests schematized in Table 1 occurred in random order. To minimize recency effects at least three items appeared between an item's occurrence in Sublist 2 and its test appearance. All nine response digits were used in each list; incorrect test pairs were composed by interchanging responses of the negative test pairs of a list.The pairs were disp...

show abstract

Section: Discussioncontrasting

confidence: 57%

Absence of within-list PI and RI in short-term recognition memory

Bower

Bostrom

1968

Psychon Sci

View full text Add to dashboard Cite

show abstract

“…The empirical findings at issue can be briefly summarized as follows. When acquisition is by a recall procedure and retention is tested by associative matching, some RI is typically found, although of a smaller order of magnitude than on an MMFR test (e.g., Delprato, 1971;Garskof, 1968;Garskof & Sandak, 1964;Sandak & Garskof, 1967). Such measures suffer from the difficulty that there is a shift in procedure between acquisition and the test of retention, which may possibly have more adverse effects on an experimental group learning two lists than on a control group learning a single list.…”

Section: Tests Of the Hypothesis Of Response-set Interferencementioning

confidence: 99%