Is the effect of reading misspellings context dependent?

In 48 university students performing single-item spelling recognition, prior exposure to misspelled words improved slightly the accuracy on correctly spelled words and increased markedly the 'false alarm' rate (classifying a misspelling seen at study as correct). In a group given a dictation test (N = 24) the only effect of exposure to misspellings was a small increment in the number of misspellings that matched the misspelling seen at study. The two test groups showed no advantage of having the same display format at study and test (AA or BB vs AB or BA). Experiment 2 (in progress) investigated a format match at study and test against a condition with a new test context (AA or BB vs AC or BC). The results to date suggest an influence of memory of the study trial rather than simply an updating by the study exposures of abstract lexical representations. [ABSTRACT FROM AUTHOR]