A comparative evaluation of parent-training interventions for families of chronic delinquents.

Fifty-five families of chronically offending delinquents were randomly assigned to parent-training treatment or to service traditionally provided by the juvenile court and community. The families in the parent-training group received an average of 44.8 hours of professional contact (23.3 hours of which were phone contacts), and each control group family received treatment estimated at more than 50 hours on the average. Comparisons of police contact data at baseline and subsequent years for the two groups showed that subjects in both groups demonstrated reduced rates of offending during the followup years. The finding most relevant was significant treatment-by-time effect for offense rates, with most of this effect accounted for by a greater reduction in serious crimes for the experimental group during the treatment year, and a similar reduction of the community control group occurring in the first of three followup years. These early decrements in offense rates persisted during followup for both groups. Throughout the study, boys in the experimental group spent significantly less time in institutional settings than did boys in the control group. Parent training had a significant impact, but the reduction in offending was produced at very high emotional cost to staff. Although it is clear that this population requires substantial treatment resources, this study underscores the need for more work on prevention.

Overt versus covert assessment of observer reliability.

This study examined the tendency of observers to make less reliable recordings of behavioral events when a calibrating observer is absent. Experienced observers in 4 different research sites coded videotapes of family interaction using the multicategory system developed at each site. 60 tapes were coded simultaneously by randomly selected observer pairs (overt reliability assessment) and 40 tapes were coded independently by 2 observers who did not know that their entries could be compared (covert assessment). Within site, intraclass correlations (ICCs) were computed separately for both forms of reliability assessment and a variety of behaviors. Overt ICCs were very high for most behaviors in all 4 systems. The corresponding covert reliabilities were significantly lower. Covert decline was conspicuous in the first 10 min of a 1-hour coding session. Hence, observer fatigue was not its principal cause. Apparently, observers lapse into a less attentive "set" prior to coding without a partner. This tendency is most discernible when a highly complex system is employed.

The utility of tracking child behavior.

To assess the impact of observation training and tracking of child behavior, each of 20 elementary school teachers participating in a behavior modification practicum was assigned to either an experimental or a control group. For each teacher, a disruptive, distractible child was identified as appropriate for behavioral intervention. All teachers were taught identical intervention strategies, but the two groups differed in the degree to which teachers were initially trained and subsequently required to observe and record discrete child behaviors. Multiple measures were used to assess the influence of monitoring child behaviors. These included observed child behavior, observed teacher behavior, teacher ratings, and correspondence between teacher perception and child behavior. Results showed that tracking the behavior of children had little or no effect on any measure.

Effects of serial dependency on the agreement between visual and statistical inference.

Comparisons between visual and time-series inferences from behavioral data show that serial dependency in scores is likely to disrupt agreement between the two methods of analysis. If researchers follow an earlier recommendation that time-series analysis be used to supplement or confirm visual analysis, this study's findings suggest that the two methods will disagree most often when the data contain high levels of autocorrelation and when reliable behavorial changes are indicated by time-series analysis.

Improving the validity of global ratings.

To examine the effects of behavior sampling on global ratings, four groups of 10 teachers each received varying amounts of observation training, practice, and feedback. Teachers viewed a series of seven videotapes depicting two, boys whose percentage of distractible behavior was systematically manipulated. Ratings of distractibility were obtained for each taped vignette. Results showed that teachers who received observation training and who routinely collected data in their own classroom submitted ratings which corresponded to actual levels of distractible behavior. Teachers who received no training, or who were trained but did not practice, submitted ratings that were significantly less accurate.