(1) Most of the specific examples of computer applications referred to in this paper were carried out under a grant from the National Science Foundation, no. 13 043.

(2) This can be seen even as early as Durkheim's work. When he was not concerned with systematic data, as in The Division of Labor, he was most fully focussed on the organization of society. When he turned to empirical investigation, in Suicide, he was least so. He was enough a sociologist that Suicide focussed on the relation of the individual to society; but it was not concerned with the very functioning of the social system itself, as was Division of Labor

(3) This example is taken from a paper by Kirk, Jerome and Coleman, James, “Interaction in a 3-Person Group”Google Scholar, mimeographed, Johns Hopkins University Department of Social Relations.

(4) Breton, Raymond, “Output Standards and Productive Behavior in Industrial Work Groups,”Google Scholar mimeographed, Johns Hopkins University, Department of Social Relations.

(5) The other was developed by William McPhee, and was used in the Wisconsin primary of the 1960 presidential election in conjunction with a survey carried out by Roper Associates. A modified version of it is reported in chapter iv in Formal Theories of Mass Behavior (Glencoe Free Press, 1963).Google Scholar The one to be described here was also used in the 1960 election, in conjunction with a panel of 800 persons in Baltimore, interviewed at points throughout the campaign. It is reported in Coleman, James and Waldorf, Frank, “Study of a Voting System with Computer Techniques”, ms.Google Scholar

(6) It might well be that the values of the parameters necessary to produce a good fit would be valuable results. These values would be those which, operationally, led the simulated system to behave as the actual system behaved, and thus would be estimates of the effects operating in the real situation. But this seems a poor means of estimating these effects, and insofar as different combinations of values would give similar results, might be misleading.

(7) The actual effect is somewhat greater than shown, because the changes over the school year, which served as the basis for the baseline curve, were themselves augmented by the effects of the friendship network.

(8) A description of this approach is presented in Coleman, James, “The Social System of the High School and the Game of Adolescence”, mimeographed, Department of Social Relations, Johns Hopkins University, 1962.Google Scholar Another example of a similar approach is given in Gullahorn, John T. and Gullahorn, Jeanne E., “Simulating Homans' Elementary Social Behavior”, paper presented at ASA meetings, 08 26–29, 1963, Los Angeles.Google Scholar