Does neighborhood size really cause the word length effect?
Dominic Guitard,Jean Saint-Aubin,Gerald Tehan,Anne Tolan
DOI: https://doi.org/10.3758/s13421-017-0761-9
2017-10-10
Abstract:In short-term serial recall, it is well-known that short words are remembered better than long words. This word length effect has been the cornerstone of the working memory model and a benchmark effect that all models of immediate memory should account for. Currently, there is no consensus as to what determines the word length effect. Jalbert and colleagues (Jalbert, Neath, Bireta, & Surprenant, 2011a; Jalbert, Neath, & Surprenant, 2011b) suggested that neighborhood size is one causal factor. In six experiments we systematically examined their suggestion. In Experiment 1, with an immediate serial recall task, multiple word lengths, and a large pool of words controlled for neighborhood size, the typical word length effect was present. In Experiments 2 and 3, with an order reconstruction task and words with either many or few neighbors, we observed the typical word length effect. In Experiment 4 we tested the hypothesis that the previous abolition of the word length effect when neighborhood size was controlled was due to a confounded factor: frequency of orthographic structure. As predicted, we reversed the word length effect when using short words with less frequent orthographic structures than the long words, as was done in both of Jalbert et al.’s studies. In Experiments 5 and 6, we again observed the typical word length effect, even if we controlled for neighborhood size and frequency of orthographic structure. Overall, the results were not consistent with the predictions of Jalbert et al. and clearly showed a large and reliable word length effect after controlling for neighborhood size.
psychology, experimental