The two following experiments were carried out in order to point out that earlier studies of social reinforcement were defective in the way of presenting reinforcement, and to state that the social context should be chosen as a determinant of social reinforcement effectiveness. The manipulations in each experiment had much in common. First, day-nursery boys and girls were subjected to a 10-minute treatment session, in which they received the reinforcing stimuli either twice (deprivation) or 16 times (satiation) from a male experimeter. This was followed by a discrimination test made of 75 trials, in which the same reinforcing stimuli as in treatment session were given to all correct responses by the experimenter, At the end of the test, they were inquired about their awareness of response-reinforcement contingencies. The measure analyzed was the number of correct responses in the test. In experiment I, the partial replication of Massari (1971) study was carried out to point out that earlier studies of social deprivation-satiation were defective in the way of presenting reinforcement. Forty subjects were instructed to read picture books in treatment session, in which they received the reinforcing stimuli, either “orikou-san-dane”(“good child” in Japanese) or a sound of bell, on the fixed schedule. This was followed by a discrimination test. The dependent measure for the four experiment groups was subjected to an analysis of variance. The results of the analysis showed no significant effects for the type of reinforcement (social-nonsocial) and the treatment (deprivation-satiation). It was discussed that the mechanical presentation of reinforcement in treatment session caused these results. And it was proposed that social reinforcement, in order to find the social deprivation-satiation relation should be presented in social context, and that the procedure in treatment session should be modified. Experiment II was carried out to show that social context was an important factor of social reinforcement effectiveness. One hundred and thirty-six subjects took part in playing with building blocks in a 10-minute treatment session, either with an experimenter (social interaction condition) or alone (no interaction condition). In this session, two-thirds of them received the stimulus words on the fixed schedule, but one-third of them received no words. This was followed by the test. In interaction condition, only boys made more correct responses in the deprivation group than in the satiation group. Among girls in interaction condition, Ss of the satiation group tended to make fewer responses than those in no word condition. These results suggested that the social deprivation-satiation hypothesis was supported only when reinforcement was given in the context of social situation, not when it was given according to the method used in earlier studies. In a general discussion, it was pointed out that social reinforcement ought to be presented in a social context.