Methods, Characteristics, Structure and Early History of
Attitude Measurement Scales

Abstract:
Measurement of attitudes is beneficial in various aspects of day to day life. Business, social and government research often rely on the measurement of respondents attitudes to guide decision and policy making. Specific research designs and methods are required to ensure useful and valid quantitative data are end results of attitude research projects. Ground breaking researchers in the field of Sociology and Psychology developed the first reliable attitude research and scaling techniques over 50 years ago that are still in use today.

An “attitude” is a theoretical entity constructed to characterize certain underlying response tendencies (Arul, 2012). As a hypothetical concept, attitudes cannot be measured directly. Any attempt to measure them “can only be inferential in nature: that is, we can only study behavior 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” (Arul, 2012, pg. 1). Attitudes have three components: affective, cognitive, and behavioral (Zikmund, Babin, Carr, Griffin, 2013). The affective part refers to person’s general feelings or sentiments toward an object. Put another way, a person’s attitudes are driven directly by their beliefs or thoughts. This cognitive component represents an individual’s knowledge about attributes and their consequences. Researchers can measure an attitude by making inferences based on the way individuals respond to questionnaires. According to Zikmund (2013), “Because we can’t directly see these phenomena, they are known as latent constructs, hypothetical constructs, or just simply constructs” (pg. 315). An attitude may include characteristics of strength, magnitude, intensity or importance. Commonly, however, “attitude measurements are concerned with the magnitude dimension and its direction; that is, the degree of favorableness or un-favorableness of a person with regard to a psychological object” (Arul, 2012, pg.1).
People have positive and negative attitudes toward various inputs and have them in varying degrees. But why should researchers study and measure these attitudes? According to Arul (2012), “Attitudes are action tendencies and as such they can facilitate or hinder action at all levels: individual, group, community, state, and national” (pg. 2). In addition,
Most managers hold the intuitive belief that changing consumers’ or employees’ attitudes toward their company or their company’s products or services is a major goal. Because modifying attitudes plays a pervasive role in developing strategies to address these goals, the measurement of attitudes is an important task…This attitude research can be directly channeled into managerial action (Zikmund, 2013, pg. 315-316).
How are attitudes measured? As mentioned previously, researchers arrive at measures of attitudes by inference. But they need data on which to base their inference. Such data are collected by various methods. However, before methods for measuring attitudes are discussed, there are some basic characteristics of measurement that should be considered in order to decide if an evaluation procedure is an effective one. According to Roberts, Laughlin and Wedell (1999),
Good tests have these characteristics. Basically, a quantitative approach to attitude measurement requires that measures (have) -
Validity: The instrument must be appropriate for what needs to be measured. In other words, a valid test measures the construct for which it is designed. A test of "attitude toward chemistry" will have items that deal directly with the concept of chemistry. Reliability: The measure should yield consistent results. In other words, if people were to take a reliable test a second time, they should obtain the same, or nearly the same, score as they got the first time they took the test, assuming no changes occurred between the two tests.
Simple to administer, explain, and understand: Generally, the measures that yield a single score of an attitude position epitomize the intent of this characteristic, although the single score may be deficient in meeting the intent of other characteristics of good measurement. Most tests of single attitudes have about 10 to 30 items, are valid, and have reliability estimates above .80.
Replicable: Someone else should be able to use the measure with a different group, or in a different situation, to measure the same attitude. Replicable tests of attitude should be usable in a variety of situations. In other words, a test of computer anxiety should measure the existence of that construct in college students, parents, elementary schools students, and even stockbrokers (pg. 211-212).
According to Roberts (1999), there are four widely used and accepted categories, or approaches, for collecting attitude information. These approaches are:
Self-reports: Where the members of a group report directly about their own attitudes. Self-reports include all procedures by which a person is asked to report on his or her own attitudes. This information can be provided orally through the use of interviews, surveys, or polls, or in written form through questionnaires, rating scales, logs, journals, or diaries. Self-reports represent the most direct type of attitude assessment and should be employed, unless the people who are being investigated are unable or unwilling to provide the necessary information. Questions like "How do you feel about ‘X’ where X is the attitude construct under investigation are often asked in self-reports.
Reports of others: Where others report about the attitudes of a person or group. When the people whose attitudes are being investigated are unable or unlikely to provide accurate information, others can be questioned using interviews, questionnaires, logs, journals, reports, or observation techniques. Parents of children can be asked how their children feel about ‘X’, where X is the attitude construct under investigation.
Socio-metric procedures: Where members of a group report about their attitudes toward one another. Socio-metrics are used when the researcher desires a picture of the patterns within a group. Members of groups can be asked questions like ‘Who in your group fits the description of ‘X’’, where X is the attitude position being studied.
Records: Which are systematic accounts of regular occurrences, such as attendance reports, sign-in sheets, library checkout records, and inventories. Records are very helpful when they contain information relevant to the attitude area in question. For example, when a researcher is trying to determine if a school wide program to develop a higher level of school pride is working, the school's maintenance records might give an index of the program's effectiveness. If school pride is improving, then vandalism should decline, and maintenance costs should be lower. The amount of trash picked up from the school's floors might yield relevant information, too. Students who have school pride are less likely to throw trash on the floor (Roberts, 1999, pg. 212-213).
Within each of these categories, there are strategies for measuring attitude-related behaviors. Most commonly, attitude measurement is accomplished by one of the following techniques:
Questionnaires and rating scales: Questionnaires and rating scales are instruments that present information to a respondent in writing and then require a written response, such as a check, a circle, a word, a sentence, or several sentences. Attitude rating scales are special kinds of questionnaires. They are developed according to strict procedures that ensure that responses can be summed to yield a single score representing one attitude. Questionnaires and rating scales are often used because they permit anonymity, permit the responder time to answer, can be given to many people simultaneously, provide uniformity across measurement situations, permit relatively easy data interpretation, and can be mailed or administered directly. Their main disadvantage is they do not permit as much flexibility as do some other techniques.
Interviews: Interviews are face-to-face meetings between two or more people in which the respondent answers questions. A survey is a highly structured interview. Often surveys are conducted over the telephone, an approximation of face-to-face interviewing. A poll is a headcount. Respondents are given a limited number of options and asked to select one. For example, word-of-mouth procedures, such as interviews, surveys, and polls, are useful because they can be read to people who cannot read or who may not understand written questions. They guarantee a relatively high response rate, they are best for some kinds of information especially when people might change their answers if responses were written, and they are very flexible. There are two major problems with interviews. First, they are very time consuming. Second, it is Possible that the interviewer may influence the respondent.
Written reports, such as logs, journals, and diaries: Logs, journals, and diaries are descriptions of activities, experiences, and feelings written during the course of the Program. Generally they are running accounts consisting of many entries prepared on an event, on a daily or weekly basis. The main advantage of this approach is that reports provide a wealth of information about a person's experiences and feelings. The main problem is in extracting, categorizing, and interpreting the information. Written reports require a great deal of time by both the respondent and the researcher.
Observations: These procedures require that a person dedicate his or her attention to the behaviors of an individual or group in a natural setting for a certain period of time. The main advantage of this approach is its increased credibility when pre-trained, disinterested, unbiased observers are used. Formal observations often bring to attention actions and attitudes that might otherwise be overlooked. Observations are extremely time consuming, and sometimes observers produce discomfort in those they are observing. The presence of an observer almost always alters what is taking place in a situation (Roberts, 1999, pg. 213-214).
A specific strategy for attitude measurement should be chosen which is suitable for the type of attitude construct of interest, the type of subject, and the situation being studied (Henerson, Morris & Fitz-Gibbon, 1978). The techniques noted above are those most often used. Others strategies are available, but researchers are cautious (or should be) when selecting a method suitable to their research questions and a method they are capable of carrying out.
Historically, the self-descriptive type of case study method is the oldest of all (Droba, 1932). In the self-descriptive case study, the individual describes orally or in writing his own attitude toward the issue in question. Case studies for various purposes were made long before questionnaire studies. Thomas and Znaniecki's “Polish Peasant in Europe and America” (1918) is one of the earliest attempts to study attitudes by a self-descriptive case study method. According to Droba (1932), “Among the later investigators who have applied the method in a more limited and probably more accurate sense Bogardus stands out as the best representative. A drawback of the case method is that it is not amenable to quantitative analysis” (pg. 312).
In order to assess the degree of attitudes possessed by people and to be able to study a large number of people, the scaling technique was introduced into attitude measurement. Various attitude measurement scales have been developed. This paper will continue on to broadly outline the characteristics of some of the earliest and most significant attitude scales so as to become familiar with the general steps involved in their structure and use.
Attitude scales provide a quantitative measurement of attitudes, opinions or values by summarizing numerical scores given by researchers to people’s responses to sets of statements exploring dimensions of an underlying theme. According to Payne and Payne (2004),
Although quantitative research is often said to be less interested in the meanings that people attach to their actions, many surveys do in fact enquire into this area. Market research in particular asks about evaluations of products and services. The main survey method used to tap meanings is attitude scaling. Attitude scales consist of asking informants to respond to a statement (or a question) in terms of a fixed range of levels... The logic behind attitude scales, drawing on social psychology, is that people are assumed to discriminate systematically in their views (Eysenck, 1953). Responding to suitable statements enables respondents to express their views. Their discriminations form a continuum from positive to negative orientations to the statements. Combinations of their discriminations can be brought together in a way that reflects underlying attitudes, which relate to other sociological variables. It is important to differentiate between a simple ‘opinion’ or ‘reaction’ to a single issue and an attitude set. Questionnaires often include questions about specific opinions, when the research is concerned with the answers themselves. Attitude scales are less interested in the specific answers, except as a means of identifying the supposed underlying attitude set. A whole range of issues can be addressed in this ‘single issue’ way. (Payne & Payne, 2004, pg.38)
Getting verbal statements from subjects requires that the subjects perform tasks like ranking, rating, sorting, or making choices. “A ranking task requires the respondent to rank order a small number of stores, brands, feelings, or objects on the basis of overall preference or some characteristic of the stimulus” (Zikmund, 2013, pg. 316). Rating requires that the respondent estimate the magnitude a certain characteristic exists. A quantitative score is the result of the respondent’s choices. According to Zikmund (2013),
The rating task involves marking a response indicating one’s position using one or more attitudinal or cognitive scales. A sorting task might present the respondent with several different concepts printed on cards and require the respondent to classify the concepts by placing the cards into groups. Another type of attitude measurement is choice between two or more alternatives. If a respondent chooses one object over another, the researcher assumes that the respondent prefers the chosen object, at least in this setting. (pg. 316-317)
Perhaps the most common practice in business research is using rating scales to measure attitudes (Zikmund, 2013). There are several different types of scales of varying complexity and purpose. According to Payne (2004) their common features are: * The presentation of a series of stimuli (usually statements); * a requirement that the response to each must be one selected from a fixed and limited choice; * the scoring of responses into a numerical value; * and some combination of these numerical scores into a single number on a scale. (pg.41) Attitude scaling requires that a respondent agree or disagree with a statement or respond to a single question. This type of self-rating scale categorizes respondents “into one of two categories, thus having only the properties of a nominal scale, and the types of mathematical analysis that may be used with this basic scale are limited” (Zikmund, 2013, pg. 317). However, simple attitude scaling can be used “when questionnaires are extremely long, when respondents have little education, or for other specific reasons. A number of simplified scales are merely checklists: A respondent indicates past experience, preference, and the like merely by checking an item. In many cases the items are adjectives that describe a particular object” (Zikmund, 2013, pg. 317). However, simple scales do not allow for fine distinctions between attitudes. Several other scales have been developed for making more precise measurements.
The phrasing and presentation of the stimulus statements are guided by the same rules that apply to good questionnaire design (Payne, 2004).
The language should be simple and avoid technical jargon. Words with loaded significance or particular meaning should be excluded, including ‘never’, ‘always’, ‘only’ and ‘almost’. Statements should be short (as a rule of thumb, not more than 20 words), each consisting of a single, uncomplicated sentence. Double negatives should be avoided. Each statement must be clear and unambiguous. It should cover a single topic. Statements need to be self-contained, dealing with only one feature, covering its aspects without overlapping into other ideas. Instructions, particularly in ‘self-completion questionnaires’, need to be clear. An example is usually given before the first statement task. Sets of statements need to be grouped together, and not run over the page, because this can sometimes confuse respondents. (Payne, 2004, pg. 38-39)
Bryman recommended that where space allows, it is better to offer the alternative in a vertical layout (Bryman, 2004, pg. 134–135). This reduces the likelihood of respondents inadvertently choosing responses that do not reflect their views, and makes coding the answers easier and faster. For example:
Islam does not separate politics and religion.
Strongly agree ___ agree ___ undecided ___ disagree ___ strongly disagree ___

Islam does not separate politics and religion. strongly agree ___ agree ___ undecided ___ disagree ___ strongly disagree ___
However, it is obvious from this example that a vertical layout takes up much more space, and there will be resource limitations to this. According to Payne (2004),
When a horizontal layout is inevitable, not all statements should take the same format, i.e. with agreements always listed on the left and disagreements on the right of the page (or coming first or last in the sequence). This is because respondents can lapse into a fixed pattern of ticking the boxes. By reversing the thrust of some statements, informants have to respond to each stimulus on its own terms (when we score such reversed statements, we also reverse the numbering system: e.g. 1 becomes 5 so that all statements then count in the same way) (Pg. 40).
This is called reverse coding and the values associated with reverse coding would be viewed as such: Old Value | New Value | 1 | 5 | 2 | 4 | 3 | 3 | 4 | 2 | 5 | 1 | Some of the earliest and most influential attitude scales developed were the Thurstone Scale, the Semantic Differential Scale and the Likert Scale. The following section will discuss some of the history behind the development of these scales, the construction of the scales and uses of these scales.
Thurstone Scale:
In psychology, the Thurstone scale was the first formal method for measuring an attitude. It was developed by Louis Leon Thurstone in 1928 as a way of measuring attitudes toward religion. The Thurstone Scale is made up of statements about the particular issue at hand, and each statement is assigned numerical value indicating how favorable or unfavorable it is judged to be (Thurston, 1929). Respondents check each of the statements they agree with, and a mean score is computed. This mean score is an indication of their attitude (Chettliar, 2011).
According to Arul (2012),
To construct the Thurstone scale a large number of statements are collected which express various possible opinions about the issue or object of study. These statements, after an editing for relevance and clarity, are given to judges, who are to independently sort them into eleven sets along a continuum that ranges from most unfavorable, through neutral, to most favorable. The eleven sets of statements are to occupy positions in the continuum in such a way that the positions are at equal intervals; that is, the difference between any two adjacent positions is the same as the one between any other two adjacent positions…For the final form of the scale, only those items are retained that have high ‘inter-judge’ agreement. When a Thurstone scale is ready, every statement in it has a numerical value already determined. When administered, the respondent just checks the items he agrees with and his attitude score is the mean value of the items he checked. (pg. 5)
According to Chettliar (2011),
The Thurstone attitude measurement procedure is generally more consistent with empirical characteristics of disagree-agree responses, it is constructed by the method of equal appearing intervals, in which a large pool of candidate statements about an attitude object, ranging from strongly negative, through neutral, to strongly positive , are sorted by a group of judges into eleven categories. They are assumed to appear equally spaced on the attitude continuum, according to how favorable the statements are towards the attitude object. Items that yield the highest level of agreement among the judges as to their scale position, and that collectively represent an adequate range of contents and scale positions, are then selected for the final scale. (pg. 1) Respondents to the scale choose only the items with which they agree, and an individual respondent's score is calculated as the mean of the items chosen. It is also called an equal-appearing interval scale (Chettliar, 2011).
According to Chettliar (2011),
Thurstone was one of the first and most productive scaling theorists. He actually invented three different methods for developing a one-dimensional scale: the method of equal-appearing intervals; the method of successive intervals; and, the method of paired comparisons. The three methods differed in how the scale values for items were constructed, but in all three cases, the resulting scale was rated the same way by respondents.
Semantic Differential Scale
Per Arul (2012),
The now-classic research by Osgood and his colleagues, based on extensive factor-analytic studies across cultures, has shown that people understand, or give meaning to, words or concepts along three dominant dimensions--the evaluative (good-bad) dimension, the potency (strong-weak) dimension, and the activity (active-passive) dimension. It has also been found that scores on the evaluative dimension correlate highly with other measures of attitude toward a particular social object. (pg. 6)
The Semantic Differential scale, developed by Osgood, Suci and Tannenbaum in 1957, can be used to measure attitudes from the meaning which people give to a word that is related to an attitude (Arul, 2012). This scale consists of a series of bipolar descriptive words such as good-bad, fair-unfair, pleasant-unpleasant, clean-dirty, happy-unhappy (etc.). In between the two words there are seven spaces. The endpoints are the opposite (or bipolar) words and the midpoint is deemed the neutral position. A sample of the bipolar continuum is given below: Fair 1_______2_______3______4______5_______6_______7 Unfair Valuable 1_______2_______3______4______5_______6_______7 Worthless Good 1_______2_______3______4______5_______6_______7 Bad
Suppose, by means of the Semantic Differential, you want to measure an individual's attitude toward The Affordable Care Act. The respondent (Joe the Plumber) is given a set of bipolar adjectives (such as the ones sampled above) and he is asked to indicate as to where for him The Affordable Care Act falls in each continuum. The numeral corresponding to the position checked by the subject is his score for that continuum. Joe’s overall attitude score is the sum (or the mean) of the scores on all the continua. The semantic differential technique originally was developed as a technique for measuring the meanings of objects or the “semantic space” of interpersonal experience (Zikmund, 2013). According to Zikmund (2013),
Researchers have found the semantic differential versatile and useful in business applications. The validity of the semantic differential depends on finding scale anchors that are semantic opposites. This can sometimes prove difficult. However, in attitude or image studies simple anchors such as very unfavorable and very favorable work well. (pg. 320-321)
Likert Scale
The Likert scale was introduced in 1932 as a scale of attitudes in Rensis Likert's "A Technique for the Measurement of Attitudes”. The Likert Scale is a bipolar scale running from one extreme through a neutral point to the opposite extreme. The Likert technique presents a set of attitude statements. Subjects are asked to express agreement or disagreement of a five-point scale. Each degree of agreement is given a numerical value from one to five. Thus a total numerical value can be calculated from all the responses (Likert, 1932).
A Likert scale is a psychometric scale commonly used in questionnaires, and is the most widely used scale in survey research (Chettliar, 2011). When responding to a Likert questionnaire item, respondents mark their level of agreement to a statement. According to Arul (2012),
For the Likert scale, various opinion statements are collected, edited and then given to a group of subjects to rate the statements on a five-point continuum: 1=strongly agree; 2=agree; 3=undecided; 4=disagree; and 5=strongly disagree. The subjects express the degree (one to five) of their personal agreement or disagreement with each of the statements. Only those items which in the analysis best differentiate the high scorers and the low scorers of the sample subjects are retained and the scale is ready for use. To measure the attitude of a given group of respondents, this scale is given to them and every respondent indicates whether s/he strongly agrees, agrees, is undecided, disagrees, or strongly disagrees with each statement. The respondent's attitude score is the sum of her/his ratings of all the statements. For this reason, the Likert scale is also known as the scale of Summated Ratings. (pg. 7)
Below is an example of a typical Likert scale questionnaire. Please select the number below that best represents how you feel about your recent online software purchase for each statement. Strongly Strongly Agree Agree Undecided Disagree Disagree ================================================================== 1. The software I wanted was easy 1 2 3 4 5 to find. 2. The checkout 1 2 3 4 5 process was easy 3. The software 1 2 3 4 5 solved my needs 4. I am happy with 1 2 3 4 5 my purchase
Source: SurveyGizmo
A Likert item is simply a statement which the respondent is asked to evaluate according to any kind of subjective or objective criteria; generally the level of agreement or disagreement is measured. Often, five ordered response levels are used (Vanek, 2012). The format of typical Likert item scales can vary. Below are common category scales:

ReferencesArul, M. J. (2012). Measurement of Attitudes. Retrieved November 12, 2013, from http://arulmj.net/atti2-b.htmlBryman, A. (2004). Social research methods. Oxford: Oxford University Press. Chettliar, C. (2011). SOCIAL PSYCHOLOGY. Retrieved November 12, 2013, from http://www.ciciliachettiar.com/social_psychologyDroba, D. D. (1932). Methods for measuring attitudes. Psychological Bulletin, 29, 309-323. doi:10.1037/h0074726Eysenck, H. J. (1953). Uses and abuses of psychology. London: Penguin Books. Henerson, M. E., Morris, L. L., Fitz-Gibbon, C. T., & University of California, L. (1978). How to measure attitudes. Beverly Hills, Calif: Sage Publications. Likert, R. (1932). A Technique for the Measurement of Attitudes. Archives of Psychology, New York University, 22(140). Payne, G., & Payne, J. (2004). Key Concepts in Social Researh. London: Sage Publications. Roberts, J. S., Laughlin, J. E., & Wedell, D. H. (1999). Validity Issues in the Likert and Thurstone Approaches to Attitude Measurement. Educational and Psychological Measurement, 211-233. Scaling: Thurston scale. (n.d.). Retrieved 16, 2013, from http://psychology.ucdavis.edu/sommerb/sommerdemo/scaling/enrich/thurstone.htmThomas, W. I., & Znaniecki, F. (1918). The Polish peasant in Europe and America. Chicago, Ill: University of Chicago Press. Thurstone, L. L., & Chave, E. J. (1929). The measurement of attitude: A psychophysical method and some experiments with a scale for measuring attitude toward the church. Chicago, Ill: The University of Chicago Press. Vanek, C. (2012, April 24). Likert Scale ? What is it? When to Use it? How to Analyze it? Retrieved November 17, 2013, from http://www.surveygizmo.com/survey-blog/likert-scale-what-is-it-how-to-analyze-it-and-when-to-use-it/Zikmund, W. G., Babin, B. J., Carr, J. C., & Griffin, M. (2013). Business research methods (9th ed.). |