College students (n=192) read two 1-1/2 page passages on the concepts of autism and trochaic meter. After reading, four independent groups answered questions requiring either a summary, the generation of a new example of the concept, a listing of the critical attributes of the concept, or the identification of a new example. All groups received feedback, and a control group received feedback but no post-question. Alternate forms of a test of concept classification were given one day and one week after reading. The identification post-question group showed greater accuracy in the classification of new examples of the autism concept than did the control group after one day, but no groups differed after one week. (Author)