of Clustering in the Recall of Randomly Arranged Associates · W. A. Bousfield et al. The Journal of Psychology. Volume 36, – Issue 1. Bousfield, W.A. BousfieldThe occurrence of clustering in the recall of randomly arranged associates. Journal of General Psychology, 49 (), pp. Psychol., 49 (), pp. Google Scholar. Bousfield et al., W.A. Bousfield, B.H. Cohen, G.A. WhitmarshAssociative clustering in the recall of words.
|Published (Last):||25 April 2012|
|PDF File Size:||6.92 Mb|
|ePub File Size:||14.39 Mb|
|Price:||Free* [*Free Regsitration Required]|
Given that the clustering scores obtained using any given model of semantic similarity are likely to be only noisy reflections of any true patterns in the data, one should use multiple models of semantic similarity whenever possible. As described below, the recall sequences are constructed to maximize semantic clustering bousfiled to g p for each participant. This word pool has been used in several published free recall studies Sederberg et al.
Across thesimulated recall sequences, and combining across the two semantic similarity measures, the observed semantic clustering scores ranged from 0. Serial effects in recall of unorganized and sequentially organized verbal material. Although the similarity values produced by each of these myriad similarity metrics are somewhat related, the pairwise correlations between the measures tend to be surprisingly low.
Another measure of semantic similarity, termed the Google similarity distance Calibrasi and Vitanyi,uses the Google search engine to compute the number of web pages containing both word x and 1593relative to the total number of pages containing each word individually; a similar metric relies on Wikipedia links to measure the similarities between topics Milne and Witten, In most free recall studies, g p is unknown. Measuring semantic clustering effects requires making assumptions about which words participants consider to be similar in meaning.
Our results provide a number of useful insights into the interpretation of semantic clustering effects in free recall.
For each recall transition we create a distribution of semantic similarity values using f between the just-recalled word and the set of studied words that have not yet been recalled.
Over the past decade, a number of techniques have been developed for systematically quantifying the relative meanings of words. Distribution of the pairwise LSA-derived semantic similarity values for the words shown in Table 1. When there is a tie, we score this as the percentile falling halfway between the two items.
Rather, we simply found the semantic clustering score toprovide a convenient means of quantifying semantic clustering. 1935 semantic analysis LSA; Landauer and Dumais, derives a set of pairwise similarity values by examining the co-occurrences of words in a large text corpus.
See other articles in PMC that cite the published article. This process continues until the k th word is recalled. Journal of Experimental Psychology. The full distributions of similarity values bousfiel from the two metrics are shown in Figure 1Panels A and B.
Cognitive Psychology and its Applications: Suppose the simulated participant has just studied a list of n words. As defined above, the semantic clustering score according to metric g p is maximized i.
Footnotes 1 Here the functions f and g p are mappings from two words, a and bonto scalar similarity values. Hippocampal and neocortical gamma oscillations predict memory formation in humans. The Google similarity distance.
There was a problem providing the content you requested
In the present manuscript we use simulations to study these questions. Generating recall sequences that maximize the semantic clustering score As defined above, the semantic clustering score according to metric g p is maximized i.
We select the word with the highest semantic similarity as the next recall, i 2and remove i 2 from the pool. Oscillatory patterns in temporal lobe reveal context reinstatement during memory search.
We 1593 by selecting the first recalled word, i 1at random from the set of n studied words. Author information Copyright and License information Disclaimer. For example, the recency and primacy effects buosfield to the well-established tendency of participants to show superior recall of items from the ends, and to a lesser extent, from the beginnings of the studied lists Deese and Kaufman, ; Murdock, The primacy, recency, and temporal clustering effects may be measured objectively by examining the relative probabilities of recalling or transitioning between items that appeared at each serial position on a studied list.
The semantic clustering score, developed by Polyn et al.
Interpreting semantic clustering effects in free recall
The order in which participants choose to recall words from a studied list of randomly selected words provides insights into how memories of bousfiekd words are represented, organized, and retrieved. Predicting human brain activity associated with the meanings of nouns.
By analyzing recall bousfisld during free recall, researchers have uncovered a number of trends that many participants exhibit. In particular, how should the magnitudes of semantic clustering effects be interpreted? The second batch of simulations used the identical set of item lists, presented to the same simulated participants. Results We ran two batches of simulations.
Interpreting semantic clustering effects in free recall.
However, for the second batch of simulations, we generated recall sequences that maximized the semantic clustering scores according to WAS-derived similarity. In this way, if a participant always chose the closest semantic associate, then their semantic clustering score would be 1. This tendency to successively recall semantically related words is termed semantic clustering Bousfield and Sedgewick, ; Bousfield, ; Cofer et al.
This indicates that bousffield semantic similarity metrics used in analyses of semantic clustering bousfiield introduce slight biases.
Bousfieldd, it is important to use multiple measures of semantic similarity if one is to obtain an accurate estimate of whether participants are semantically clustering their recalls. We next generate a percentile score by comparing the semantic similarity value corresponding to the next item in the recall sequence with the rest of the distribution.
We generated 5-item recall sequences that maximized the WAS-derived semantic clustering score forsimulated participants presented with 50 item lists each see text for details.
Table 1 Simulation word pool.