One observer coded interactions at a microanalytic level while the
other independently used a rating scale (i.e., JAFA) to score the interactions holistically.
Interrater reliability estimates were calculated for all measures by research assistants who
were unaware of the questions of the study