From our discussion of the nature of attitudes in an earlier chapter, we know that attitude is an hypothetical entity constructed to represent certain underlying response tendencies. As hypothetical constructs, attitudes cannot be measured directly. Any attempt to assess them can only be inferential in nature: that is, we can only study behaviour which is reasonably assumed to indicate the attitudes to be measured and quantify these indications so as to get an idea of how much individuals or groups differ in their psychological orientations toward a particular object or issue. The behaviour to be studied for the purpose may be one that occurs in a natural setting or a simulated situation; it may be speech or performance behaviour. For example, a person talking about, say, prohibition when he is seated at a teastall over his usual cup of tea exhibits speech behaviour in a natural setting. If you were to formally interview him on his views about prohibition, you would again have his speech behaviour, but in an artificial environment. Observation of his visits to bars or drink sessions at home would provide data on his natural performance behaviour. If, on the other hand, he served as a subject in an experiment designed to study his behaviour, we would obtain performance data from a simulated setting.
An attitude may comprise characteristics of strength, magnitude or intensity, importance, salience or centrality, complexity, flexibility, etc. Commonly, however, attitude measurements are concerned with the magnitude dimension and its direction; that is, the degree of favourableness or unfavourableness of a person with regard to a psychological object. A psychological object, as we have already seen in the chapter on nature of attitudes, may be a person, a group, an idea, a symbol, or anything with regard to which people could harbour positive or negative feelings.
Why measure attitudes
People have likes and dislikes and have them in varying degrees. But why study and measure them? Attitudes are action tendencies and as such they can facilitate or hinder action at all levels--individual, group, community, state, and national.
Consider the case of population control. Among the various possible activities toward progress, the government of India wanted its people to adopt birth control and compulsory sterilization was introduced in certain parts of the country. But, this move, as we all know, gave rise to immense anxiety and antagonism among people and the programme ended in a fiasco. Such blatant disregard for people's attitudes not only thwarted many a corresponding programme of the government, but, tagged on to other grievances of the kind under Emergency, also turned into a massive timebomb which, after an adequate gestation period, blew up the once invincible reign of Mrs. Gandhi and her colleagues. At the same time, certain other activities like the introduction of new variety seeds in agriculture and the introduction of adult education had a better fate. To determine, therefore, what action to introduce when and how to introduce it for the desired effect among a target population, the action planners must know how far the existing action tendencies of the population are receptive/resistent to the proposed action. Such knowledge would help devise appropriate means for triggering the desired change.
It is not being suggested here that attitudes have to be changed first in order that intended behaviour may follow; nor is the reverse being implied that behaviour must be changed so that corresponding attitudes will emerge. In fact, attitude may influence behaviour and in turn be influenced by it. Whatever the theoretical position on this what-comes-first issue, it is a fact that attempts to introduce a change, more often than not, face some degree of resistence. Such resistance is dependant on how favourably or unfavourably the population in question is inclined toward the proposed course of action and related issues. It is here that knowledge of the target population's attitudes can guide any effort for planning and implementing developmental change. Action plans which ignore the attitudes of people can at best look forward to chance success.
It is indeed likely that action change, once introduced and adopted, gives rise to a host of new attitudes towards the action and other connected issues in life. People's attitude toward themselves, having been exposed to a new course of action, may undergo change. They may also develop positive or negative attitudes toward the change agents. Such emergent attitudes, if of the unfavourable kind, will come in the way of future efforts. Assessment of the psychological impacts of interventions is, therefore, of interest not just to psychologists. All those engaged in any kind of persuasive, promotional, or developmental activities would need similar measures both in order to judge the expediency of introducing a particular action and to know the effects of having introduced it. Further, measurement of attitudes is a necessity for experimental and other studies concerned with attitude change.
Approaches to Attitude Measurement
How do we measure attitudes? We have said that we arrive at measures of attitudes by inference. But we need data on which to base our inference. Such data are collected by various methods. We may observe the ongoing behaviour of people in the natural setting; we may directly ask the respondents to state their feelings with regard to the issue under study; we may assign a well-defined task to respondents and record their performance; we could generate data by giving the respondents partially structured stimuli to interpret or react to. Physiological reactions of respondents when exposed to the attitudinal object can also provide us with relevant data. For example, Westie and DeFleur (1959) studied the attitudes of subjects by taking the latter's galvanic skin responses, heartbeat, etc, when presented with pictures of Negroes and Whites in various social situations. Hess and Polt (1960) found that pleasant stimuli gave rise to pupillary dilation, whereas unpleasant ones led to pupillary constriction on the part of subjects. There is yet another set of methods of data collection, known as unobtrusive or non-reactive methods, in which respondents' co-operation or involvement is dispensed with. Let's elaborate on these methods:
This method involves recording the actual behaviour of people whose attitude is to be studied. It is indeed an objective method and well suited for certain kinds of issues. For example, it is quite commendable to observe the actual overt behaviour of strikers by participating in the strike itself to gain a measure of the strikers' attitude. We could also observe a company executive in his day-to-day dealings with his subordinates to assess his attitudes towards them. Not all issues, however, lend themselves to direct observation. Can you, for instance, use this method of data collection to study the attitude of voters? Even the most dedicated and non-partisan psychologist would not have access to the polling booths to observe the actual voting behaviour of people. Again, if you wanted to study and measure attitudes towards sex, could you possibly observe the overt sexual behaviour of people? Some situations may, no doubt, permit simulation of a natural setting and role-play can be enacted to be able to observe behaviour and infer the underlying attitude(s).
Even where accessibility is not a problem, direct observation of behaviour is not practicable if we want to have data on a large number of individuals. In addition to the difficulty of observing every individual in a large sample, there is no telling as to when an individual will exhibit the behaviour which is relevant to the study in question. Thus, over and above the intense effort and cost involved, indefinite patience may also be called upon. Further, even when we have managed to spot some behaviour related to our study, we do not know if the behaviour was an outcome of the related attitude or one caused by other factors. Take for instance a boy who goes to church regularly. You have observed his behaviour, all right. But, does this behaviour mean that he is favourably inclined toward prayer and religion? Not necessarily. On asking him directly, you may discover that his girl friend cherishes religious sentiments and attends prayer services regularly. He goes there only to meet her!
Observation of behaviour, even when the behaviour is the outcome of the attitude being studied, may tell us the direction of the underlying attitude (i.e., whether it is positive or negative), but it cannot as easily indicate the magnitude or strength of the attitude. It is also difficult to establish the reliability of behavioural measures of attitudes. The observers' perception as well as their ability to report what they have observed vary considerably. All these weaknesses, however, do not suggest that observation of overt behaviour is futile for assessing attitudes. They only caution us that care needs to be taken in deciding upon the suitability of this method to a given situation.
If we want to know how people feel about a certain thing, it seems most natural to ask them straight away as to what their feelings are. Direct questioning has, therefore, had an appeal as a method of studying attitudes. But, however logical and smooth this technique may seem to be, it serves only a limited purpose of roughly classifying respondents as favourable, unfavourable, and indifferent with regard to a psychological object. Here we have the same problem as the one often encountered in direct observation; normally, neither direct observation nor direct questioning assesses the degree of attitudes an individual possesses. In the absence of such an index, we cannot discriminate among individuals within the favourable and the unfavourable groups; nor can we know the distance between the two groups. There may, of course, be times when we require only to know as to how many are "for" and how many "against" a particular issue. Direct questioning may adequately serve the purpose here.
We must bear in mind that inhibiting and/or social desirability factors can contaminate responses to direct questioning. For example, if you were to study people's attitudes toward the National Emergency when the Emergency was on, many a respondent would be reluctant to give you an answer or would, in all likelihood, loudly proclaim a favourable attitude. When controversial issues are involved and pressures and threats are operative, direct questioning is not the suitable means of data collection for assessing attitudes.
Even when no threats are present, not all individuals are capable of articulating their feelings. A person may possess certain attitudes and behave accordingly, but may not be aware of them. Psycho-analysts have borne abundant witness to such phenomena. Thus, direct questioning or any other self-report technique will be of little avail if the respondent has no access to his own attitudinal orientations, buried in the realms of the unconscious. We shall return to the drawbacks of self-report methods a little later.
Some other approaches
Projective techniques and objective tasks, where the attitude objects are disguised, are successful ways of overcoming many of the difficulties encountered in direct observation. But these methods are not very popularly used in measuring attitudes -- probably because of low reliability in interpreting such data. Measures of galvanic skin responses (GSR), pupil dilations, vascular constrictions, heart-beat, etc. can also indicate attitudes. But such physiological arousals may not always discriminate attitudes. For instance, both positive and negative tendencies may give rise to similar readings in the galvanic skin response apparatus. Physiological measures have, no doubt, been carefully and gainfully used on certain occasions and interested readers may refer to Leiderman and Shapior's (1964) Physiological Approaches to Social Behaviour.
All the different methods used for measuring attitudes have their own strengths and limitations. One method may be more suitable than another for a given study and the same method may be inappropriate on another occasion. There is no one method, therefore, which is universally good or well-suited. However, going through the literature on attitude measurement, we find that self-report methods have been used most frequently. Techniques other than self-report have been tried and their applicability has been proven and advocated. But in practice, the pencil-and-paper type of self-report measures have stolen the scene to this day.
Scales of Measurement
To have more refined measures of attitudes than rough classifications of for and against, some scaling devices can be made use of. Before going into scaler measures of attitudes, we shall first discuss scales of measurement in general.
Measurement is assignment of mathematical symbols to objects and events according to rules. In order to assign different symbols to different objects, one must be able to differentiate objects on a given aspect, attribute or property. Such differentiation may be rough and crude or may be refined and specific. You may, for instance, want just to classify objects, persons, or responses into different categories. A nominal scale will suffice for this purpose. The only criterion to assign "objects" to different categories of a nominal scale is whether the objects are the same or different with regard to the property being studied. To classify individuals, for example, according to the province they come from or according to the religion they belong to would constitute a nominal scale. If you assign numbers to the different categories in this scale, the numbers are just identification names. They are not amenable to mathematical operations like calculations of means, coefficients of correlation, etc. You can, of course, count the number of subjects under each category lable (numeral or verbal) and find the modal category in which the highest number of individuals fall. You may also perform a test of association, if you categorised the individuals according to two (or more) attributes. For example, if you categorised individuals both according to their province and their religion, you could perform the Chi-square test to see if a particular province(s) tend(s) to be associated with a particular religion.
If you want to know the relative positions of persons or objects with respect to a characteristic, you need an ordinal scale, in which individuals or objects are ranked as first, second, third, etc., depending on the more or less of the attribute possessed by the individuals or objects. The ordinal scale can state who has more or less of the attribute under study, but not how much more or how much less. If person P is ranked first, Q second, R third, etc., we cannot know if the difference between P and Q is or is not the same as the difference between, for instance, U and V; the magnitude of difference between any two consecutive ranks remains unknown and is likely to vary.
An interval scale can tell us whether P is as much higher than Q as Y is than Z on a particular attribute. In other words, in an interval scale, the difference between any two adjacent positions is the same as the one between any other two adjacent positions. Thus, the interval scale is an improvement over the ordinal scale, even as the latter is over the nominal scale.
There is another type of scale, called the ratio scale, which is commonly used in the physical sciences. To have a ratio scale the absolute zero point needs to be determined. A ten-inch rod can be said to be exactly twice as long as a five-inch one, because both the rods share a common starting point, namely, the real zero point. But in the subject matter of the social sciences, the zero point is arbitrary and, therefore, we cannot express relationships in terms of strict ratios. Psychophysics has made attempts, in limited areas, to establish absolute zero points. By and large, however, social sciences do not use ratio scales; they employ ordinal and interval scales in their studies.