Nonreinforced imitation in children: effects of prior reinforcement histories
Date
1974
Authors
Oliver, Peter Roger
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
In three experiments, the effects of prior reinforcement for complying or not complying with adult instructions were assessed with respect to the amount of nonreinforced imitation which occurred in a subsequent modelling situation. Kindergarten children were first reinforced for following or not following a series of instructions in a nonimitative setting. The occurrence of nonreinforced imitation was subsequently assessed in a dual-model paradigm in which one experimenter (SD model) modelled behaviours and rewarded the child for imitating while a second experimenter (S-delta model) modelled both similar and dissimilar behaviours but never reinforced the child for imitating. In all cases, the S-delta model was the same experimenter who had previously reinforced the children for compliance or noncompliance. The prediction was tested that children who had previously been rein forced for following adult instructions (Compliance Group) would show larger amounts of nonreinforced imitation than children who had no prior experience with the S- delta model (Control Group). Also tested was the prediction that children who had previously been reinforced for not following adult instructions (Noncompliance Group) would show lesser occurrences of nonreinforced imitation than children in the control group.
In Exp. I, one group of children were reinforced for complying with adult instructions prior to the modelling situation while a control group received no prior experience with the S-delta model. As predicted, the amount of nonreinforced imitation was higher following compliance training than in the control group.
Control Ss in Exp. I were not given any prior experience with the S-delta model as had the children in the compliance group. Consequently, the increased amount of nonreinforced imitation following compliance training could have been due to prior experience with a rewarding model per se. To preclude this possibility, Exp. II directly replicated the procedures of Exp. I with the addition of a third condition in which a separate group of children were first reinforced for not complying with the same instructions given to the compliance group. Since children in this noncompliance group received approximately the same amount of reinforcement as children in the compliance group, it served as a control for the effects of prior rewarding experiences per seas well as a test of whether nonreinforced imitation could be decreased following noncompliance training.
Exp. II successfully replicated the results of Exp. I. Children previously reinforced for following instructions displayed significantly more nonreinforced imitation than both control Ss and children given the same amount of rewarding experience with the S-delta model but for noncompliant behaviour. Children in the noncompliance group imitated fewer behaviours without reinforcement than control Ss but this difference was not statistically significant. Lack of a significant decrease appeared to be associated with sex differences with boys requiring additional training procedures than those used with girls in order to establish noncompliant behaviour. These differences in training noncompliance could have affected nonreinforced imitation for boys to a greater extent than for girls.
Exp. III replicated the procedures of Exp. II while strengthening and standardizing noncompliance training for both boys and girls. Exp. III also varied the sex of the S-del ta model so that an equal number of boys and girls were exposed to both male and female models. Prior reinforcement for compliance again increased the occurrence of nonreinforced imitation, successfully replicating the results of Exps. I and II. Noncompliance training also resulted in a significant decrease in nonreinforced imitation compared to both compliance and control groups. These effects were independent of the sex of the child and/or the sex of the S-delta model.