Contrary to expectations, none of the interactions was significant. Failure to replicate previous studies was attributed to control of examiner bias, superior statistical and experimental controls, and to differences in subject characteristics. The efficacy of modeling films was not conclusively demonstrated. (Author)