INTEROBSERVER AGREEMENT IN ANALYZING CONFLICT RESOLUTION SITUATIONS INVOLVING YOUNG CHILDREN.

Abstract

The study was designed to determine interobserver agreement in analyzing behavioral conflict interactions of young children. Selected portions of recorded interactions were subjected to four observers. An observational instrument containing 26 well defined observational categories was used by observers to analyze interactions. Prior to the use of this observational instrument, it was pilot tested and necessary improvements were instituted. Observers were provided systematic training for a total period of 14 hours. At the termination of the training period each observer viewed the selected videotape episodes with the help of a prepared word-by-word transcript. Observers were to mark the occurrence of behavioral interactions on a specifically devised record form. Recorded interactions from each observer were analyzed by use of Cohen's Kappa for determination of interobserver agreement. The findings indicated that obtained K's ranged from a low of .60 to a high of .69 suggesting a moderate reliability of observational instrument. A related finding suggested that observers perception of difficulty involved in using the behavioral categories was directly related to interobserver agreement for categories with certain levels of difficulty. Implications of these findings for replication were discussed.