Rules and Self-Rules: Effects of Variation upon Behavioral Sensitivity to Change
Baumann, Ana A., Abreu-Rodrigues, Josele, Souza, Alessandra da Silva, The Psychological Record
Verbal statements describing contingencies of reinforcement or punishment, such as instructions, rules, advice, or commands, have been most widely denominated as "rules" (Skinner, 1969). Rules have important effects on behavior. Under some circumstances, rules may facilitate the development of contingency control. For example, in the presence of rules, contingency-appropriated behavior tends to be acquired faster (e.g., Ayllon & Azrin, 1964). This effect has been reported when contingencies are complex (Baron, Kaufman, & Stauber, 1969; Takahashi & Shimakura, 1998), imprecise (Cerutti, 1991), or aversive (Galizio, 1979). Under other circumstances, however, rules may impair control by the prevailing contingencies. Accordingly, studies have shown that when contingencies change, such that rules no longer correspond to the current contingency, rule-controlled behavior may remain unaltered (e.g., Kaufman, Baron, & Kopp, 1966; Paracampo, Souza, Matos, & Albuquerque, 2001; Shimoff, Catania, & Matthews, 1981). When behavior, verbally mediated or not, does not change with changing contingencies, it can be said that behavioral insensitivity was observed. If, on the other hand, behavior changes when contingencies change, it is said that behavior was sensitive to the change in contingencies (Madden, Chase, & Joyce, 1998).
Because much of human behavior is taught through rules, and behavioral insensitivity may not always be adaptive, many studies from human operant laboratories have focused on identifying the variables of rule-controlled behavior that increase behavior sensitivity to change. Sensitivity to change is more likely to occur when (a) the prevailing contingencies do not support rule-governed behavior (e.g., Galizio, 1979; Newman, Hemmes, Buffington, & Andreopoulos, 1994), (b) rules do not fully describe task performance and/or contingencies (e.g., Danforth, Chase, Dolan, & Joyce, 1990; Raia, Shillingford, Miller, & Baier, 2000), (c) the new task is of low complexity (Albuquerque & Ferreira, 2001), (d) individuals have a history of extinction or punishment for following rules (e.g., Hackenberg & Joker, 1994; Martinez-Sanchez & Ribes-Inesta, 1996), and (e) competitive social reinforcement for rule following (i.e., a pliance contingency) is not present (e.g., Barrett, Deitz, Gaydos, & Quinn, 1987; Hayes, Brownstein, Zettle, Rosenfarb, & Korn, 1986; Navarick, 2004; Otto, Torgrud, & Holborn, 1999).
Another factor that promotes sensitivity of rule-controlled behavior is a history with varied rules and schedules. In a study conducted by LeFrancois, Chase, and Joyce (1988), undergraduate students were assigned to one of three groups: variety rule, specific rule, and specific control. During the training phase, the variety rule group was exposed to three types of schedules of reinforcement: variable interval (VI), fixed time (FT), and fixed ratio (FR). Half of the participants were also exposed to a differential reinforcement of low rates (DRL) schedule. The specific rule and specific control groups were exposed to only one type of schedule: variable ratio (VR) or VI. Accurate schedule descriptions (hereafter called rules) were presented to the participants in the variety rule and specific rule groups before each reinforcement schedule. Participants in the specific control groups did not receive rules. During the testing phase, in which no rules were given, a fixed-interval (FI) schedule was in effect. In the training phase, low, moderate, and high response rates were observed for the participants in the variety rule group. High response rates tended to be presented by participants in the specific rule and specific control groups exposed to the VR schedule, whereas low rates were observed for participants in the specific rule group exposed to the VI schedule. In the testing phase, response rates decreased with the introduction of the FI schedule for the participants in the variety group, while response rates remained unaltered for participants in the specific rule and specific control groups. …