Scaling and sociometric techniques

Scaling and Sociometric Techniques

A scale is a set of symbols or numerals so constructed that the symbols or numerals can be assigned by rule to the individuals (or their behaviors) to whom the scale is applied, the assignment is indicated by the individual's possession of whatever the scale is supposed to measure. As a test, a scale is a measuring instrument.

Scaling

Scaling is the branch of measurement that involves the construction of an instrument that associates qualitative constructs with quantitative metric units. Scaling evolved out of efforts in psychology and education to measure unmeasurable constructs such as authoritarianism and self-esteem. In many ways, scaling remains one of the most arcane and misunderstood aspects of social research measurement. It attempts to do one of the most difficult of research tasks-measure abstract concepts.

Social Scaling

Scales differ from indexes in that they take into account some qualities about the nature of the relationship of the separate items to each other. This is referred to as the intensity structure among the items. Because of this additional consideration, scales are more complex to devise. Scales are not easy to develop, and they may create more error in measurement than a less complex instrument. Remember that the most important aspect of developing a scale is to create items that measure what you are trying to measure (items that have face validity). In addition, these items should cover the various domains of content that the scale is purporting to represent; that is to say, the issue of content validity is an overriding one in scale development.

There is another option: use an already constructed scale. We will end with a discussion of the types of scales available to measure occupational status. Even when you select an already constructed scale, you should nevertheless understand how it was made. If you do not, you are likely to use it inaccurately and may find interpreting your results very difficult. The scale types discussed here are not all equivalent to one another. Some refer largely to how the variation in answers is set up. Still, others are distinguished by the responses given.

1. Likert-type (summated rating) scales

Rensis Likert (1932) has developed a technique for helping to eliminate questionable items from the scale. This is a summated scale consisting of a series of items to which the subject responds. The respondent indicates agreement or disagreement with each item on an intensity scale. The Likert technique produces an ordinal scale that generally requires non-parametric statistics. Perhaps the most widely used form of scaling in survey research sets up ordinal categories for degrees of agreement, generally including the five levels of "strongly agree," "agree" "disagree," "strongly disagree," and "don't know" (or "undecided"). These response categories are attached to a set of statements. Assuming that the responses to each statement are equivalent, you can assign scores of 1 through 5 (or 0 through 4) and can create an index by summing the scores and averaging them. Because the items are added up, a Likert scale is in some ways an index of items that consistently scaled response categories.

If there are many statements in the set, you may want to condense them in some way. One procedure is to create a total index of the items and then carry out an item analysis to see which items are most closely related to the index; you might then use these to form a smaller index. A more complex procedure is to carry out a factor analysis a statistical technique, which examines all the interrelationships between the various items to determine which sets of items are most strongly related. These sets of items are called factors; they serve as a new dimension of the concept being measured.

2. Semantic differential scale

The Semantic Differential (SD) method was developed by Osgood, Suci, and Tannenbaum in 1957. A Semantic Differential Scale offers bipolar positions (such as two adjectives: "active/ passive") to a single stimulus (for example: "my mother"). The respondent is asked to rate the stimulus on a seven-point (or five-point) scale. Usually, as in the above examples, the bipolar items are adjectives; the stimulus, a reference to a person or persons. To review this, the bipolar adjectives might be "strong-weak," "happy-unhappy," "tense-relaxed," and the stimulus "I am." Somewhere between the bipolar extremes of these adjectives would commonly be a seven-point scale on which the subject must place the stimulus referent at a point, which seems most appropriate.

The reason why the Semantic Differential Scale is useful is that a respondent may not always be sure how to describe in words exactly how intelligent or not intelligent, how timid or bold, or how happy or unhappy another person is. Placing the person on a scale running from one extreme to the other is easier. In some cases, instead of adjectives, brief statements are used at one end, their converse at the other. Furthermore, a study may contain more than one semantic differential, each with the same set of bipolar items, but with different subjects to apply them to. Then the ratings on each item can be compared.

3. Bogardus social distance scales

This is a scale that focuses on the distances between the ordinal items in the scale (Bogardus, 1959). It has been widely used to measure the attitudes toward ethnic groups, but it has also been employed to measure attitudes toward occupational, social class, and religious groups. It asks the respondent to think of the group to be rated in terms of the type of social interaction in which he or she would choose to engage with members of that group. The forms of social interaction vary in their degree of intimacy.

Total scores on cumulative scales have the same meaning. Given a person's total score, it is possible to estimate which items were answered positively and negatively. A pioneering scale of this type was the scalogram. Scalogram, Scalogram analysis is a procedure dimensional if the responses fall into a pattern in which endorsement of the item reflecting the extreme position results also in endorsing all less extreme items.

4. Thurstone scales

Thurstone Scales consist of several items whose position on the scale has been determined previously by a ranking operation performed by judges. The subject selects the responses that best describe how he feels. Thurstone Scales are composed of items selected by judges as indicative of measuring some concept. The general procedure is that the researcher would amass a large number of items seemingly measuring a particular concept, for example, conservatism. Judges would be asked to classify each item, independently, on a scale of 1 to 11, from the strongest measure of conservatism to the weakest. Once the panel of judges has completed the classifying, the researcher determines and averages scores for each item from the average responses of the judges.

This means that the items selected relate to one another in such a way that the ones to which the judges gave higher scores presume agreement with the items given lower scores. Thurstonescaling is one of the earliest and best-known forms of scaling. However, it is very time-consuming, and it has been challenged by some methodologists as very problem-ridden. For these reasons, it is not often used today.

Consensus scaling requires items to be selected by a panel of judges and then evaluated on 

(1) relevance to the topic area, 

(2) potential for ambiguity, and 

(3) the level of attitude they represent. 

A widely known form of this approach is the Thurstone equal-appearing interval scale. Also known as the Thurstone scale, this approach resulted in an interval rating scale for attitude measurement.

These scale values, of course, are not shown on the questionnaire, and the items are usually arranged in random order rather than in order of their scale value. The mean (or median) of the scale values of the items the individual checks is interpreted as indicating that individual's position on a scale of favorable attitude toward the object.

Guttman'sscalogram analysis is cumulative in the sense that the combination of responses required to make a particular score includes the responses to all questions required to make the next lower score, plus the response to one additional question, in a stepwise fashion. For example, suppose that we wish to walk up a stairway containing 10 steps. We number all steps from 1 to 10 beginning at the bottom. Now imagine that we have all of our respondents walk up as many steps as they wish and stop. They may not skip any steps (step up more than one step at a time). We will give them 1 point each for each step they traverse. It should be clear that there is only one way to make each score. To make a score of 3 a person must walk up steps 1,2, 3 only. To make a score of 8 he or she must walk up steps 1,2,3,4, 5,6,7,8 only. We say the scale is cumulative because a score of 4 is made by walking up 3 steps for a score of 3, plus the fourth step; a score of 8 is made by walking up the first 7 steps (for a score of 7) plus the eighth step. Since each score represents a unique set of responses, we know from a person's score exactly which steps he or she walked up (or exactly with which items he or she agreed).

Sociometry Techniques

Sociometry is a technique for describing the social relationship among individuals in a group. Indirectly, it attempts to describe attractions or repulsion between individuals by asking them to indicate whom they would choose or reject in various situations. Children in a school classroom may be asked to name in order of preference (usually two or three) the child or children that they would invite to a party, eat lunch with, sit next to, work on a class project with, or have as a close friend. Although some researchers object to the method, it is also common to ask the children to name the children, again in order of preference, that they would least like to invite to a party, eat lunch with, sit next to, and so forth.

As a means of presenting simply and graphically the entire structure of relations existing at a given time among members of a given group. The major lines of communication, or the pattern of attraction and rejection in its full scope, are made readily comprehensive at a glance (Hall, 1948: 11).

Origin of Sociometry

The basic principle and techniques of sociometry were first embodied in a volume entitled Who shall survive? J.L. Moreno, first published in 1934. An enlarged revised edition of this book published in 1953 covers the history, theory, terminology, techniques bibliography, and applications of sociometry.

Sociometry Choice

The basic technique in sociometry is the sociometric test. This is a test under which each member of a group is asked to choose from all other members those with whom he prefers to associate in a specific situation. The situation must be a real one to the group under study, e. g ., group study, play, and classroom seating for students of a public school.

A specific number of choices say two or three to be allowed is determined concerning the size of the group, and different levels of preferences are designated for reach choice. The sociometric choice may mean the choice of persons, choice of lines of communication, or choice of lines of influence.

Applications of Sociometry

Sociometric techniques are widely used by sociologists, psychologists, and psychiatrists in the study of group structure, status, personality traits, leadership, morale, and social adjustment. They make more explicit and precise the configuration of group relationships, the characteristics and composition of cliques and other elements in a large group, the social position of individual members in a group, and the streams and points of influence within and among groups. Sociometric studies of communities, fraternities, schools, college student bodies, camps, armed forces, and factories have been made.