Two versions of a simulation game, Parent-Child, were taught to ten white and ten black inner-city males. The twenty subjects played a total of 198 rounds of the game, and the following conclusions were made with respect to the use of the Parent-Child game as a research site. (1) Inner-city fifth and sixth grade males can be taught to play Parent-Child. (2) The subjects play the game less effectively than a totally rational player would, but give some behavioral indication of understanding how one should play the game in order to win. (3) The subjects' behaviors are role and structure specific, which allows the simulation game to be used as a research site in a larger study of the effect of role and structural constraints on game behavior. (Author)