Some support for expressed concern over the negative consequences of computer game play was obtained by Colwell and Payne (2000) in a questionnaire survey of UK adolescents. A second study was carried out in Tokyo, Japan, to test for generality of findings. There was strong evidence of similarity in relation to play patterns, but in contrast to the UK sample, number of good friends and self-esteem did not relate to game play. A principal components analysis of a scale to measure needs met by game play resulted in the same two previously obtained factors; 'companionship' and 'prefer to friends'. However, in contrast to the UK sample, 'companionship' did not relate to play. A positive relation between the 'prefer to friends' factor and play was obtained for both boys and girls (boys only in the UK), thus providing support for the 'electronic friendship' hypothesis. There was evidence of a positive relation between play and aggression, but a preference for aggressive games was associated with lower aggression scores, and this raises questions for the 'causal hypothesis'.