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):||16 January 2006|
|PDF File Size:||3.90 Mb|
|ePub File Size:||11.90 Mb|
|Price:||Free* [*Free Regsitration Required]|
We also found that the semantic clustering scores computed using LSA were slightly bousfisld reliably higher than those computed using WAS paired t -test: If these bousfifld objective semantic similarity metrics based on huge text corpora and experimental datasets fail to agreeon a set of pairwise semantic similarities, how could one possibly expect to study effects of semantic organization in individual participants?
One might accomplish this by using, for example, an LSA-derived scoring model while using a WAS-derived internal similarity model in their simulation or vice versa as we have done here.
There was a problem providing the content you requested
Rather, a near-ceiling clustering score may reflect the specific sequence of words presented to the participant, or the specific structure of the experiment. The content is solely the responsibility of the authors and does not necessarily represent the official views of our supporting organizations.
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. If one observes or fails to observe a similar pattern of clustering scores across bouwfield conditions when using multiple semantic similarity models e.
The order in which participants choose to recall words from a studied list of randomly selected words provides insights into how memories of the words are represented, organized, and obusfield. Because this procedure ensures that each recall will be followed by the most similar word bouusfield is yet to be recalled, by definition it will maximize the semantic clustering score according to g p.
Our simulations are intended to estimate the maximum expected magnitude of semantic clustering effects in free recall. Cognitive Psychology and its Applications: Introduction The free recall paradigm has participants study lists of items — typically words — and subsequently recall the studied items in the order they come to mind. Word association spaces for predicting semantic similarity effects in episodic memory.
The dotted gray lines indicate the means of each distribution. As defined above, the semantic clustering score according to metric g p is maximized i. Distribution of the pairwise WAS-derived semantic similarity values for the same words.
We ran bousfielr batches of simulations. We first divided the 1935 of LSA-derived pairwise similarity values into equally sized bins the centers of the bins are plotted along the x -coordinate. Over the past decade, a number of techniques have been developed for systematically quantifying the relative meanings of words. We then measure the degree of semantic clustering according to a different similarity metric, f.
A neurosemantic theory of concrete noun representation based on underlying brain codes. Weobtain a single bohsfield clustering score for each simulated participant by averaging the semantic clustering scores across all lists that the participant encountered.
We then create 195 pool of the n – 1 remaining words from the studied list. Table 1 Simulation word pool. As described below, the recall sequences are constructed to maximize semantic clustering according to g p for each participant.
Given that the clustering scores obtained using any given model of semantic similarity are likely buosfield be only noisy bousfie,d of any true patterns in the data, one should use multiple models of semantic similarity whenever possible.
Latent 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. Two measures of semantic similarity A. This tendency to successively recall semantically related words is termed semantic clustering Bousfield and Sedgewick, ; Bousfield, ; Cofer et al.
Predicting clustering from semantic structure. Measuring semantic clustering effects requires making assumptions about which words participants consider to be similar in meaning. In particular, how should the magnitudes of semantic clustering effects be interpreted?
Interpreting semantic clustering effects in free recall
We found that the mean semantic clustering score was 0. Semantic clustering score The semantic clustering score, developed by Polyn nousfield al. Generating recall sequences that maximize the semantic clustering score As defined above, the semantic clustering score according to metric g p is maximized i. This indicates bousfueld different semantic similarity metrics used in analyses of semantic clustering may introduce slight biases.
This panel is identical to panel E, but here we generated recall sequences that maximized the LSA-derived semantic clustering scores, and plot the distribution of observed mean WAS-derived clustering scores.
The same 5, randomly chosen item lists were used in both panels. First, 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.
One pervasive finding is that when a pair of semantically related words e. The latent semantic analysis theory of acquisition, induction, and representation of knowledge.
Results We ran two batches of simulations. Using fMRI brain activation to identify cognitive states associated with perception of tools and dwellings. Gamma oscillations distinguish true from false memories. This process continues until the k th word is recalled.
For this reason the precise clustering score one observes is difficult to interpret, and one would be better served by instead comparing distributions of clustering scores obtained across conditions in an experiment or across participants. This panel shows a binned variant of the scatterplot in panel C.
Interpreting semantic clustering effects in free recall.
We then computed the similarity between each pair of words by measuring the cosine of the angle between the corresponding LSA vectors. In most free recall studies, g p is unknown. An analysis of sequences of restricted associative responses. By analyzing recall sequences during free recall, researchers have uncovered a number of trends that many participants exhibit.
However, for the second batch of simulations, we generated recall sequences that maximized the semantic clustering scores according to WAS-derived similarity. Specifically, we calculate the proportion of the possible similarity values that the observed value is greater than, since strong semantic clustering will cause the observed similarity values to be larger than average.
University of Pennsylvania; Philadelphia, PA: Predicting human brain activity associated with the meanings of nouns. See other bojsfield in PMC that cite the published article.