The purpose of this paper is to describe a social influence process that is not vulnerable to the objections raised by psychologists and political scientists to what they call „excessive sociologizing“ in the explanation of choices.
The choices here happen to be votes — although the process to be described is not limited to this application. What is meant by „social influence“ is the observation that people in close contact for long periods, such as husbands and wives or parents and children, are found to have remarkably nonindependent preferences. At any given time, their choices are highly correlated. What is meant by objectionable „sociologizing“ of this phenomenon however, are interpretations of how these correlations come about over time. It is too easy to make the choice itself a function of deliberate social manipulation by others or conscious adaptation to others.
No one really advocates this kind of „social determinism,“ however. Nearly everyone admits the necessity, and most of us the primacy, of two other determinants of the immediate choice: namely, (1) external political stimuli such as current events and (2) internal psychological dispositions such as political party loyalties learned in the past. But if we admit the primacy of these brute facts, then how can responses to them nevertheless be so correlated with the subtle cues from others in the contiguous social environment?
This paper describes a social influence process that has the desired subtlety. It leaves the determination of the choice at all times to external (that is, political) stimuli and individual (that is, psychological) dispositions. Working through these, however, such a process is capable of generating any of the social correlations found in choice data.
2. A three-process model of voting
The influence process to be described is one of three processes that make up a model for the study of voting behavior. It is no gecident that the other two processes concern the primary determinants of voting referred to above: the response to external stimuli and the learning of internal dispositions. The complete model has been described in detail elsewhere. It was designed and is being used for large problems of national scope and historical time spans. Therefore, the processes by which a single individual forms a single vote in one election are necessarily too simple to be interesting as psychological theories. Here they will be briefly explained and with a certain amount of literary license by the flow diagram of Figure 1.
This model is operated in a computer. We shall assume that a large sample of „voters,“ representing a community in some political period, has already been put into the machine storage or memory of the computer. Each voter is represented by a limited series of characteristics, many of which will be modified by events during each election period. Let us assume that we are about to start an „election,“ as one of a sequence of them.
The model first takes from the input hopper an iniection of „political stimuli“ for the period since the last election. These are distributions of stimuli graded as to their attractiveness or cogency from strong to weak. The distributions differ for different groups in the sample electorate, as determined by needs of the problem not relevant here.