" . . . subjects were given an Initial List of three very favorable trait-names, while [other] . . . subjects were given a different Initial List consisting of three very unfavorable trait-names. Each subject was instructed to form an impression of a fictitious person, and then to select from a check list four other trait-names appropriate to the initial impression . . ‥ Trait-names selected under the Favorable condition were consistently and considerably more favorable than those selected under the Unfavorable condition. In addition, under the Unfavorable condition the average value of selected trait-names was consistently more favorable than the average value of the associated Unfavorable Initial List. Under the Favorable condition, however, an analogous deviation toward unfavorableness was not statistically reliable." From Psyc Abstracts 36:04:4GD93P. (PsycINFO Database Record (c) 2006 APA, all rights reserved)