The typical levels-of-processing experiment uses incidental learning. A subject is asked to rate words based on the number of letters or consonants, or on the words' pleasantness. Because the subject is unaware that there will be a memory test later, the experimenter can assume that, once the rating task is over, the subject will not process the item further. After all of the ratings are done, the subject receives a surprise recall or recognition test. The usual finding is that the deeper the level of processing, the better the performance on the test. This demonstration is adapted from one reported by Craik and Tulving (1975). Ideally, you would be unaware that there will be a memory test at the end and unaware that the three tasks are designed to induce different levels of processing.