Reactivity in the classroom observation of hyperactive children

Reactivity is a critical problem in observational research, yet traditional approaches have produced inconclusive, if not contradictory, data. A dual process model, of nonspecific and focal reactivity, is proposed. Nonspecific reactivity refers to the self-conscious behavior subjects exhibit when they orient or attend to an observer. Focal reactivity refers to impression management strategies adopted by subjects with the intent of creating favorable or unfavorable impressions in observers. The present study attempted to investigate the extent of focal reactivity in the classroom observation of hyperactive children. Four classes of causal factors can be identified yet naturalistic technologies have differed in the type and extent of dependent measure confounding which has been produced. A technology for investigating focal reactivity was proposed and involved eliciting intentional response distortions which were compared with measurements made under baseline or typical conditions. Four carefully selected hyperactive and matched normal first grade boys were chosen for this study. Beginning with four days of baseline observation, all subjects were videotaped for one hour on eight consecutive school days. Following the baseline period, teachers were asked to alternately produce two-day samples of target children's "best" and "worst" behavior. As expected, measures of social behavior were most vulnerable to focal confounding while motor activity measures were generally nonreactive. Although there were differences between dependent measures and subject groups, the estimates of focal response distortions typically reflected a "fake good" response set during normal observation. Attentional behavior patterns also appeared inflated during baseline conditions, particularly for the hyperactive sample. Teachers used different tactics to influence the two types of children, though they were more successful with the hyperactives. A heightened sensitivity to the misbehavior of target children during baseline observations may have mediated teachers' efforts to create a favorable impression in observers. These results support the view that dependent variable confounding produced by teacher-mediated, focal reactivity is a significant obstacle to observational research. Specific implications and recommendations for future naturalistic studies are outlined.