INDEXING SYSTEM

A simple letter detection task, in which subjects mark instances of a target letter in prose passages, has elucidated numerous cognitive processes involved in reading by examining the "missing letter" effect, in which readers' detections accuracy is especially low on frequents function words. Two experiments explore the fundamental but novel issue of whether the missing letter effect is due to the test containing the target letter or to the words surrounding the test word. College students searched for a target letter (e in Experiment 1, o in Experiment 2) in a passage included unrelated sentences, with each sentence containing a single instance of 1 of 2 test words (the or one in Experiment 1, of or on in Experiment 2). The sentences were intact (prose), or the words in each sentences were randomly rearranged (scrambled). The 2 test words in an experiment were surrounded by the exact same words. If the word containing the target letter is primarily responsible for the missing letter effect, the proportion of correct letter detection responses should depend on the test word, whereas if the surrounding words are primarily responsible, it should depend on the text type (prose, scrambled). In fact, in both experiments a huge effect of test words was found but no effect of text type. These results provide clear evidence for the influence of the test word but, surprisingly, no evidence for thr influence of thr surrounding words on the missing letter effect in the letter detection task.