Abstract
In this experiment subjects studied two lists of words. One list was presented in a normal form and the other list was presented in an inverted form. Then subjects were given a word fragment completion test, in which half of the words of each list in the study phase presented in the normal form and the other half were presented in the inverted form. There were two word fragment completion tests, a fast test and a slow test. In the fast test, larger priming was obtained when the presented form of words between study and test was matched, as compared with not matched. But in the slow test, priming wasn't influenced by the presented form of words. Word fragment completion had been classified as a data-driven test, but the slow test was relying on data-driven processing less than the fast one. Considering these findings, it was concluded that word fragment completion test involves not only data-driven but conceptually driven processing, and that one can manipulate test performance so that it is data-driven or conceptually driven under some conditions.
