Cane hundred and twenty eight university students took part in a study of personality impressions. They observed one of four target persons (TP's) interacting with someone the TP's purportedly liked or disliked. The Tl''s varied the patterns of their gazing while holding total eye contact constant to see if fewer but longer glances were decoded differently than more but shorter glances. The gaze pattern of more but shorter glances produced impressions of higher activity and higher potency. The evaluation factor of the personality ratings was only affected by the interaction context with TP's appearing more positive when interacting with friends. The interaction context also affected subjects' willingness to attribute their impressions to the TP's personality rather than to the situation across both the evaluation and the activity dimensions. Cross-cultural implications of the results for gaze patterns were discussed and suggestions made for exploring the effects of various interaction contexts on personality attributions.