BINARY‐CHOICE SEQUENCES OF RETARDATES, NORMAL CHILDREN, AND COLLEGE STUDENTS UNDER RANDOM‐ AND PATTERN‐SET INSTRUCTIONS 1

Twenty successive free binary choices (between A and B ) were elicited without reinforcement from adolescent educable mental retardates, from normal fourth‐ and fifth‐grade children, and from college students under one of two sets of instructions: instructions to respond randomly and instructions to respond with a simple pattern of successive choices. There was a preference for A rather than B as the first response. For all treatment combinations, the number of alternations over the entire set of 20 items was above chance; consistent with previous findings, however, the college students alternated less than the others. Although the instructional variable had no significant effects on retardate behavior, random‐set instructions elicited very few cyclic sequences of choices from normal children and college students. College students who gave cyclic sequences generally gave longer cycles than did S s in the other two populations; most of the children who gave cyclic sequences alternated; and some of the retardates who gave cyclic sequences alternated and some perseverated. These results imply that the tendency to alternate above chance that has been noted in binary‐choice tasks is due to something more than preconceptions about the nature of random sequences. There is a general tendency to alternate that is at least as strong under pattern‐set as under random‐set instructions; the latter instructions appear to be interpreted as instructions to behave unpredictably.