Reading is a highly practiced skill that eventually becomes seemingly effortless. Indeed, the well-known
Stroop effectopens in new window
is based on the principle that, in adults, reading is automatic. Although reading an individual word feels like it happens "immediately," a large literature shows that the speed of word processing is influenced by statistical properties such as word frequency and word length.
These results are obtained from lexical decision tasks, in which subjects judge whether a letter string is a word or non-word as quickly and accurately as possible. Non-words are created by changing letters in real words to produce letter strings that are pronounceable but do not sound like real words. Despite the fact that data are collected for both word and non-word trials, researchers typically only analyze responses to words. However, understanding the mechanisms driving non-word performance could provide valuable insights into early lexical processing.
In a recent paper published in the Journal of Experimental Psychology: Learning, Memory, and Cognition, Yap and colleagues (2015) utilized the large dataset available as part of the English Lexicon Project, which includes lexical decision data from 1,200 subjects collected across 6 universities, to investigate the psycholinguistic and individual differences variables that influence non-word performance.
Psycholinguistic factors included number of letters, orthographic distinctiveness (e.g., the number of words that can be obtained by changing a single letter), average frequency of words that could be obtained by changing a single letter (base-word frequency), number of syllables, and number of morpheme-like elements (affixes). These factors accounted for 40% of the variance in non-word reaction times (for comparison, analogous factors accounted for approximately 60% of the variance in reaction times to words).
Number of letters was the strongest predictor of non-word performance, with slower and less accurate responses to longer non-words. Performance was also slower and/or less accurate for non-words that were less orthographically distinct, and had more syllables and affixes. However, these inhibitory effects decreased as base-word frequency increased. That is to say, non-words with higher base-word frequency were less influenced by the other psycholinguistic factors. This interaction with base-word frequency is qualitatively similar to results obtained with words, suggesting that common lexical processes are engaged for word/non-word discrimination.
Individual differences analyses revealed high within- and between-session reliability, indicating that performance with non-words reflects meaningful and stable individual variation, not just measurement noise. In addition, participants with high vocabulary knowledge were faster and more accurate at rejecting non-words than participants with low vocabulary knowledge. Moreover, participants with higher vocabulary knowledge were less influenced by the number of letters, but were more influenced by orthographic distinctiveness and the number of syllables and affixes, and benefited more when base-word frequency was high.
Together, these results demonstrate that non-word performance is not random noise, but rather is systematically mediated by psycholinguistic and individual differences factors, providing new constraints that can be used to refine computational models of word recognition and lexical processing.
Citation:
Yap, M. J., Sibley, D. E., Balota, D. A., Ratcliff, R., & Rueckl, J. (2015). Responding to nonwords in the lexical decision task: Insights from the English Lexicon Project.
Journal of Experimental Psychology: Learning, Memory, and Cognition
,
41
(3), 597–613.
https://dx.doi.org/10.1037/xlm0000064
Note: This article is in the Basic / Experimental Psychology topic area. View more articles in the Basic / Experimental Psychology topic area.