The Scientific Study of Social Behaviour (Psychology Revivals)

Michael Argyle

• Publisher: Routledge
• Year: 2013
• eBook ISBN: 9781135039905
• Topic: Psychology
• Subtopic: Applied Psychology
• Index: Psychology

Originally published in 1957, this book presented an up-to-date account of psychological research into human social behaviour of the time. There are chapters on interaction between pairs of people, behaviour in small social groups, and human relations in industry. The author avoided the adoption of any particular theoretical position, and concentrated on the established empirical findings of the time. The results of several hundred investigations are summarised and compared, so that the principal generalisations which emerge can be seen. Stress is placed on rigorous methods of research, and a critical account is given of current techniques of social research, showing the importance of experimental and statistical methods. Careful consideration is given to the danger of the investigator disturbing what is being investigated. Use is made of recent ideas about theory and explanation, and the different kinds of theory used in experimental psychology were considered for the first time as possible ways of accounting for group behaviour.

This book was intended not only for students of psychology and of the other social sciences, but also for industrialists, administrators and indeed all who were interested in the laws underlying social behaviour. Today it can be read and enjoyed in its historical context.

Part I

METHODOLOGY

CHAPTER II

METHODS OF STUDYING SOCIAL BEHAVIOUR

INTRODUCTION

THE STUDENT OF social behaviour is fortunate in many ways: he does not need to travel to Polynesia or to the bottom of the sea, he does not require animal houses or complex apparatus. Nevertheless the development of a science of social behaviour has been held up through a complete lack of adequate techniques. Some people still have doubts as to whether social behaviour is really capable of being treated scientifically. It is true that it raises a number of unique problems, though probably no more than any other branch of science. In this chapter these difficulties will be indicated, together with the attempts which have been made to overcome them. The position taken here is that it is just as possible to find valid empirical laws in the social sciences as it is in physics and chemistry. In the following pages a concise account will be given of the methods of research as they have been perfected so far, with particular reference to their logical features.

There are perhaps two groups of people who will disagree with this approach—plain sceptics, and exponents of ‘Action Research’.

Many people outside social science are rather sceptical of what it can do, and they sometimes maintain that it is not possible to obtain valid empirical laws in this field. Their argument is supported by reference to the special difficulties of research here—the influence of the investigator on the behaviour studied, the difficulty of social experimentation, the intermediacy of the human instrument in measurement, etc. These and other questions are dealt with in this chapter.

By ‘Action Research’ is meant any investigation in which social change or therapy is the prior object, and the discovery of scientific results a subordinate one. From a belief in the greater importance of social change, together with an awareness of the difficulty of obtaining objective empirical laws, some writers go on to say that the scientific aim should, in effect, be abandoned. “… the experiment tends to submerge its classical objective, the search for proof, and becomes instead a miniature form of social action” (Madge 1953, p. 287). This is to regard objective empirical laws as a fetish of merely academic interest. The point is that they are of direct and universal application, whereas immediate results are not. The Action Researcher ought to obtain objective results of at least two kinds: firstly, he should prove that what he is doing is genuinely effective— for example, in increasing output, reducing inter-group hostility, or curing neurotics. Secondly, he should show the precise conditions under which these successful results can be obtained, so that others can do the same. It is not suggested that he should go further and take an interest in testing theoretical hypotheses, for these are largely the concern of the research worker.

MEASURING INSTRUMENTS

The student of social behaviour has two measuring instruments which are particularly his own, though he may borrow the tools of others. These two are the different kinds of interview and questionnaire, and the techniques of controlled observation. The questions about the logic of measurement, the disturbance created by the investigator, and the design of the research will be discussed in later sections of the chapter.

Interview and Questionnaire

Modern psychologists are more concerned with behaviour than with states of consciousness. Nevertheless one kind of behaviour is verbal behaviour, and what a person says or writes is of great interest to the psychologist in several ways. One way in which the interview or questionnaire may be used is for obtaining knowledge about objective facts known to the informant but not to the investigator. For example, an anthropologist who arrived too late for a rare ceremony, such as a coronation, in a society which kept no records would have to use this method. Likewise an investigator of behaviour which he was not allowed to observe, such as sexual behaviour, would use this approach (cf. Kinsey et al. 1948). It is however familiar to the experimental psychologist that perceptions are liable to selection and distortion according to interest and mood, and that memory continues this process. If any more direct method is available, it is preferable to the interviewing of informants. Another use of the interview and questionnaire is in personality assessment. From a person's answers to certain questions and his general behaviour in the test situation, predictions are made about his behaviour in other contexts. The accuracy of this method of assessment is discussed later (p. 95 ff.). A third use of the interview is to change a person's behaviour, as in psychotherapy and social case-work. This will be discussed later as a form of social interaction (p. 112 f.), but is not a means of measurement and will not be pursued here further.

The most important use of verbal methods in the study of social behaviour is for the analysis of perceptions, attitudes, and motivation, since this cannot be carried out so well in any other way. By perception is meant the sum total of material open to introspection, about some part of a person's present physical environment, and which can be obtained from subsequent verbal reports. For social psychology the term is used in a somewhat broader sense than in experimental psychology (see p. 95). In general the best way of investigating perceptions is with open-ended questions, but it is possible to use closed multiple-response questions or rating scales. Two investigators, for instance, have measured ‘perception of the attitudes of others to oneself by this means: Horowitz and others (1951) used a graphic scale, and Pepitone (1950) a set of verbal categories. Another method of assessing perceptions is to compare the situation or stimulus with a number of standard stimuli, as in classical psychophysics. This method has not yet been used in the study of social perception proper—the perception of social objects— though Bruner and Goodman (1947) used it in their investigation of the influence of economic background on perception of the size of coins. It could be used in social perception if one had a graded series of pictures, recordings, or actual personalities against which to match the person being judged.

Attitudes are patterns of behaviour in response to particular people or objects, or classes of these. The most familiar kinds of attitude are political, religious, and inter-racial ones, though attitudes towards specific individuals can also be measured. An attitude resembles a personality trait, save that it refers to a narrower class of responses, and as in the case of traits the class of behaviours comprising the pattern is defined by the correlation together¹ of these responses over a large number of people. An important logical difference between perceptions and attitudes is that attitudes refer primarily to overt behaviour, while perceptions do not. Thus verbal measures of attitudes can be validated against behaviour, but this is not the case with perceptions. Similarly {{beliefs are defined as verbal assent to certain propositions, there is no further criterion beyond interview-questionnaire data.

The third kind of empirical variable which can be measured by interview and questionnaire is motivation. This differs from perception in that its operational definition does not rest in verbal reports, and from attitudes in that, while it may be defined in terms of behavioural criteria, these are not yet agreed upon. Motivational states may be defined as (a) the result of a certain period of deprivation, as in experiments on animals, (b) biological states of the organism, in the case of primary drives, or (c) certain kinds of behaviour, such as on the ‘activity wheel’ used for rats. What is required is research on deprived humans, after the model of the Minnesota starvation study (Keys et al. 1945), to find the concomitants of motivational states and to develop convenient means of testing for these states. The measurements could take the form of records of behaviour in certain situations, or of some kind of interview or questionnaire. One such method which has been used is assessment of fantasy by content analysis of Thematic Apperception Test protocols (McClelland et al. 1953). A second is to offer choices of activity to the subject/as in the Allport-Vernon values questionnaire (1931). A third is the use of self-rating scales such as Swanson used (1951), with questions like “How much does it mean to you to be one of the leading members of the group?”.

There are a number of types of interview and questionnaire, the more important of which will be described here, together with brief notes on their advantages and disadvantages for measuring perceptions, attitudes, and motivation.

Many kinds of information can be obtained either by interview or by questionnaire—the questions are spoken in one case and written in the other. (The unstructured and psychoanalytic types of interview cannot of course be given in questionnaire form.) When an interviewer obtains answers to a number of definite questions it is called a schedule. The questionnaire has the great advantage of anonymity, making for more truthful answers, though it was found in the course of investigations in the U.S. Army that Negro soldiers thought that their answers would be read by a white audience (Hyman 1951, p. 209), indicating that ‘audience effects’ are not always eliminated in the questionnaire. The more standard situation of the questionnaire also serves to cut out uncontrolled personal influences, and there is less likelihood of bias in the coding of replies.

The interview, on the other hand, although suffering from the corresponding disadvantages, is in general more flexible and can produce more illuminating results. It has been shown, for example, that the same question can have different meanings to different people (Crutchfield and Gorden 1947), and such misunderstandings should be avoided. The interviewer can probe for the true significance of replies, and make ratings based on the whole of the subject's behaviour. He can vary the order of the questions, and prevent the subject looking over the whole list before answering.

The following five types of interview and questionnaire will be discussed below: the social-survey interview or questionnaire, the non-directive interview, the psychoanalytic interview, attitude scales, and sociometry.

(1) Social surveys are generally conducted by a standardised interview, since not everyone who receives a questionnaire under these conditions will return it; a high-percentage return is required if the results are to be at all representative, whereas, for example, 10% to 40 % is typical for a postal ballot.

This kind of interview is necessarily short, since a large number¹ have to be completed, and consequently it consists of a series of direct questions, often with a limited range of possible answers. The use of this kind of question has been criticised by McNemar (1946) on the grounds that the answers are very unstable and liable to be influenced by such peripheral factors as emotionally charged words, the order of the questions, and so forth. Furthermore, the use of fixed alternatives may force an answer before one has crystallised, and direct questions have a way of suggesting their own answers, even when this would not be expected (Muscio 1916). However, most of these difficulties may be countered in some way—by careful pre-testing of the questions, by ‘filter’ questions to ascertain if the respondent possesses the knowledge or opinion in question, and by the inclusion of a number of open-ended questions permitting a free answer in the subject's own words. A great deal is known about the best ways of wording the questions; this is discussed by Cannell and Kahn (1953) and Cantril (1944). In carrying out a social survey it is usual to train the interviewers so that they all adopt a quite standardised approach, and variation due to the interviewer is eliminated.

(2) The non-directive interview is the opposite of the above method, and allows spontaneous conversation on the part of the subject, with the minimum of questioning on the part of the interviewer. The method was first used in the Hawthorne investigations (Roethlisberger and Dickson 1939, pp. 272–86), and was developed as a method of therapy by Carl Rogers (1942). Rogers has devised skills of interviewing which enable the interviewer to elicit answers by putting extremely vague questions, by asking for fuller information, and by making encouraging noises. This method avoids the suggestion of answers to the subject and the misinterpretation of questions, allows him to reply in his own words instead of confining him to set alternatives, and encourages him to say what seems important to him; it seeks for his reasons and will reach areas of personal importance (‘depth’). (Krech and Crutchfield 1948, pp. 276–88.)

There are two drawbacks to this method of interviewing—the subject may fail to talk about the subjects in which the interviewer is interested, and the results are difficult to score. The first difficulty may be overcome by means of semi-structured questions, so that increasingly direct questions may be asked in a ‘funnel structure’ to obtain answers on the required subjects. In the ‘focused interview’ developed by Merton and Kendall (1946), semi-structured questions are used in relation to a specified range of material to be covered, while attempts are made to discover what precise experiences and feelings are being reported. The difficulty of scoring may be overcome by the interviewer making ratings on the spot—of the extent to which certain attitudes, for example, are displayed; or he can take full notes which are afterwards rated by judges or subjected to a formal content analysis, along lines to be described later. This method is valuable for measuring drives, attitudes, and perceptions both quantitively and qualitatively; its length, however, restricts it to purely research purposes or to the pilot stage of a survey.

(3) The psychoanalytic interview is the now quite familiar process wherein the patient lies on a couch and free-associates with the help of the analyst, usually starting from memories of childhood and the recall of dreams. As will be mentioned later in this chapter, the patient is more highly motivated towards frank disclosure than are most subjects in social research, and he may become aware of material that was initially not open to conscious introspection. The method aims to disclose unconscious motivations and memories. It is claimed by psychoanalysts that this method gives results unobtainable by any other means, and that they are of greater importance for understanding a person's motivations. There are, however, grounds for comparing this method unfavourably with other kinds of interview. Firstly, no attempt is made to ask questions which are either standard (like the questionnaire and schedule) or free from suggestiveness (like the non-directive interview). Secondly, no attempt is made to obtain objective records of what is said, and it is clear that the analyst is selective in what he records, while his interpretation (which is the basis for later questions) is based on a particular theory. Thirdly, there is a somewhat complex social relationship between analyst and patient, and this may be expected to lead to particular kinds of concealment, invention, and distortion. As will be seen later, the relation between interviewer and subject is a matter of great importance. Furthermore, it is difficult to validate the psychoanalytic interview, for the following reasons: (i) The material obtained is not normally used to predict future behaviour, as for example are attitude scales, though there is no reason in principle why this method of validation could not be used, (ii) there is usually no way of...

Table of contents

• Cover
• Half Title
• Title Page
• Copyright Page
• Original Title Page
• Original Copyright Page
• Dedication
• Table of Contents
• Figures
• Preface
• Acknowledgments
• I. Introduction
• Part I. Methodology
• Part II. Generalisations and Theories
• References
• Author Index
• Subject index