Attitudes
Bookmark this page

11.1 Attitudes

11.1.1 Function and Measurement Structure and function of attitudes

Cognitive, affective and conative components of attitudes.

Functional approach: adaptive, knowledge, ego-expressive functions.

Measuring attitudes

Likert Scales,

semantic differential and projective techniques for measuring attitudes.

11.1.2 Attitudes and behaviour

Principle of consistency between attitudes and behaviour.

Attitude change

Cognitive dissonance and attitude change.

Other explanations for attitude change, including, persuasive communication and dual-process models of persuasion.

11.1.1 Function and Measurement / Structure and function of attitudes

An attitude cannot be seen or directly measured, this is in comparison to many phenomena in the scenes of biology or chemistry; as a consequence social psychologists have enveloped many definitions, different approaches to understanding and a range of measurements of attitudes.  The term attitude is used to represent quite complex mental processes.  Attitudes are important since in virtually a;; aspects of our social lives we continually seek to discover other people’s attitudes, tell others of our views, and try to change another person’s opinions.  Attitudes are also regarded as important influence on our behaviour.

Petty and Caioppo (1986) defined attitudes as:

general evaluations people make about themselves, others, objects or issues

and went on to say:

attitudes have a past, present and future; there were developed from past experience, they guide our current behaviour, and can direct our development inn the future

An attitude is evaluative and enduring  It is a relatively stable disposition to evaluate some object or event. 

The structural approach regards attitudes as an evaluation (positive or negative) of an attitude object (person or issue) and may be broken down into three components:

·        Cognitive Component - beliefs or thoughts about the attitudinal object. 

·        Affective Component - feelings about the attitudinal object. 

·        Conative Component - response or action tendencies toward object of attitude – how we intend to behave toward the person or issue. 

The functional approach to attitudes, suggested by psychodynamic psychologists such as Katz(1960)  suggests that the well being of an individual is promoted by attitudes  serving four functions:

·        Knowledge function (mainly the belief components) -  the world can be a confusing and strange place; attitudes help us to  organise our social world, thuds making the world more familiar and predictable.

·        adaptive attitudes / need satisfaction / utilitarian: attitudes help avoid "bad things" and help approach "good things.".  To do with the extent to which attitudes enable a person to achieve a desired goal or avoid undesirable things.

·        Ego defence: blaming and putting others down as regulation of self-regard, scapegoating – function clearly relates to the psychoanalytic perspective of Freud.

·        Ego- expressive Social identity: you are in part what you believe in and like; attitudes make people bond or clash. Allows us to tell others our opinions and views.  As social beings we need to communicate with others

The basic idea behind the functional approach is that attitudes help a person to mediate between their own inner needs (expression, defence) and the outside world (social an information)

Martin (1987) – investigating how children remembered gender behaviour

design  repeated measures

methods

1.      children were shown videos of people who behaved in either a gender stereotypes way or in a non-stereotypical way.

2.      children were asked to recall behaviours of both stereotypes

results

·        children tended to distort their memories to fit in with the stereotypic attitudes of the male or  female gender role

comment

these results were then due to argue that gender stereotypes help the child understand masculine and feminine in a simple way

thus demonstrating the knowledge function of attitudes

The structurakl and the functional approaches to understanding attitudes are like two sides of the same coin.  Neither apporach on its own provides a full picture.  Which approach you focus on depends on what interest you have in attitudes.  For example,  much of the measurement of atitudes has focussed on theaffective component from the structural approach.  This is because it is simple to mesure, gives a good summary of an attitude and is often a good predictor of behaviour.  By contrast, thefunctinal apprioaach may be important of you aare ineresteed in trimg tpo change attitiueds.  To change a person’s attitudes the approach should match the function, for example an attitude serving a knowledge function is most likely to be changed bby showing the person the information.

Cross-cultural research on attitudes has shown that quit general differnees between cultures may have specific consequenses for changing people’s attitudes.  For example, Hofstede (1980) studied over forty countried and tried to  locate each one on an ‘individualim-colectivism’  dimension.  A country showing individualism reflects individuals who prize personal choice and achievement. By contrast a country showing collectiism reflects individuals who value group harmony and collective achievement highly.  Hofstede, found Americans to show individulaism and people in Asian countries to show collectivism.

Han and Shavitt (1993) – to determine  whether advertisements in different countries appealed to individualism or collectivism.

design – content analysis

methods

·        analysis of American advertisements

·        analysis of Korean advertisements

results

·        American advertisements  promoted personal success and achievement

·        Korean advertisements had strong themes of group harmony

·        Americans were more likely to be influenced by an advertisement promoting personal achievement rather than group harmony

Measuring attitudes

Because attitudes are dispositions, social perceivers as well as psychologists must infer attitudes.

Social psychologist Mark Stasson's comment, "We are manipulators, not measurers," is true: social psychologists spend much more of their time thinking of ways to manipulate independent variables-- designing the various experimental treatments, examining threats to the study's internal validity, considering ways to maximize the power of their experimental manipulations--than in working out how to measure the dependent variables.

Jumping the gap between the conceptual variable and the operational definition is, perhaps, the essence of science. Many disciplines spin theories, but only sciences seek data to test these theories. Indeed, in the history of science it has been the discipline's success in measuring phenomena that has determined, in large part, its success in extending its theories.

Reliability describes the consistency of the measure across time and across its components. If the questionnaire you take today indicates that you have are prejudiced against Danierians, when you retake the test next week it should again indicate that you dislike Danierians. Also, your answers across the 10 questions should be relatively consistent; if you are prejudiced you should agree with most of the negative worded items ("I think Danierians are untrustworthy") and disagree with most of the positively worded items ("I wish more Danierians lived in my neighbourhood").

Validity, in contrast, describes the extent to which the technique measures what it is supposed to measure. To be valid, the indexes we design must actually measure the conceptual variable of interest. Validity is greatest when the link between the conceptual definition of the interpersonal or personal process and the operational definition of that process is clear.

The theoretical definition gives a conceptual description of the term. Prejudice, for example, is "the rejection of individuals based on their membership in a social category."

The operational definition, in contrast, specifies the measurement operations that are used to quantify the concept. Prejudice, operationally defined, is giving more money to the members of one's own group, rating another group negatively on a checklist, or refusing to be photographed with members of the other group. If the operation used does not measure the variable of theoretical interest, then validity is low. A questionnaire measure of prejudice, for example, may not be valid if subjects routinely misrepresent their thoughts, feelings, and intentions or they simply have difficulty describing them.

Validity can also be undermined by the social desirability bias; the tendency to answer questions in a socially acceptable way. If respondents think that prejudice is an embarrassing personal flaw, they might claim to be more open-minded than they actually are.

Self-report measures

Researchers are often interested in subjects' own views of their personal thoughts, feelings, and actions. These self-report methods, despite their variations, are all based on a simple premise: if you want to know what someone is thinking, feeling, or planning, then just ask him or her to report that information to you directly. Opinion polls often ask respondents to state their political beliefs or describe their stance on contemporary social issues. During interviews the researcher records the respondent's answer to various questions, but questionnaires ask respondents to record their answers themselves.

Self-report methods have been used for at least a century by psychologists, but their "birthdate" for social psychology is often set at 1928, when Louis L. Thurstone published an article entitled "Attitudes can be measured." To assess individuals' attitudes, he argued, simply record their answers to one or more questions pertaining to the object. With this simple suggestion, Thurstone laid the foundation for all self-report measures of attitudes. Thurstone chose to use a fixed-response approach rather than an open-ended, or free-response approach.

With free-response, or open-ended, questions, respondents must describe their personal viewpoints in their own words. Questions such as "What do you think was the main cause of the Brixton riots " and "What are the most important problems facing our coiuntry today?" are both free-response questions.

Fixed-responses items, in contrast, offer respondents a choice between two or more alternative answers. "Are you in favor of, or opposed to, the use of busing to achieve racial integration?" and "What soft-drink to you prefer: Coke or Pepsi?" are examples of fixed-response questions.

He also opted for a questionnarie rather than interview. A questionnaire asks a series of questions, which responds answer in writing. In interviews, the questions are asked by an interviewer who must transcribe the interviewees responses. Interviews are most useful when subjects may not be motivated to respond, and a good interviewer can extract valuable information from subjects. Interviewers can, however, bias the respondents' reactions.

Thurstone also opted for multiple questions rather than single questions. Opinion pollsters often measure an attitude with a single, well-chosen question, but researchers prefer to use attitude scales: a series of questions pertaining to a single attitudinal topic. Prejudice may be too complicated to assess by just asking "Are you prejudiced against Derligians? By including many items researchers can also be more confident that the wording of the items isn't biasing their results. When the items are selected and pretested for accuracy, the measure is usually termed a test or a scale. By asking about attitudes with several questions, researchers can be more confident that the wording of their questions isn't biasing their results. Also, many attitudes are so complex that several items are needed to give a full description of the respondent's perspective. Sexism, for example, may be too complex an attitude to assess by just asking "Are you in favour of, or opposed to, equal rights for women?"

Thurstone first generated a large number of statements pertaining to the attitude topic. Then he had raters (judges) rate each one's value, as reflecting a positive or negative or neutral attitude. He then selected items to create a scale that had several items for each gradation of attitude to the topic--some very positive items, some less positive items, some positive items, and so on. Subjects then responded to only those items that reflected their attitudes.

Sample items from a Thurstone scale are shown below. An actual scale would include 10 to 15 items and the scale values (shown in the parentheses) wouldn't be included.

Directions. Put a check mark in the blank if you agree with the item.

1.      Blacks should be considered the lowest class of human beings. (scale value = 0.9)

2.      Blacks and whites must be kept apart in all social affairs where they might be taken as equals. (scale value = 3.2)

3.   I am not interested in how blacks rate socially. (scale value = 5.4)

4. A refusal to accept blacks is not based on any fact of nature, but on a prejudice which should be overcome. (scale value = 7.9)

5. I believe that blacks deserve the same social privledges as whites. (scale value = 10.3)

________________________________________________________________

   (Source: Thurstone, 1931)

Another scaling method, which actually predates Thurstone's technique, was based on the concept of social distance: the distances individuals prefer to maintain between themselves and members of other ethnic groups. First suggested by E. S. Bogardus (1925), the Social Distance Scale proved to be a useful measure of prejudice, and was the forebearer of a more general, measurement method known as a scalogram or Guttman scale (Guttman, 1950). A Social Distance Scale is shown below.

An example of a Bogardus Social Distance Scale.

Directions. If you willingly admit members of a group in the specifified situation, place a check mark in the blank under the group label.

Source: Bogardus, 1925)

A third way of scaling attitudes was recommended by Rensis Likert (1932). With a Likert scale the need for independent raters is eliminated by asking respondents to indicate the intensity of their attitudes themselves. As shown below, on a Likert attitude scale subjects typically indicate degree of agreement with the item on a scale ranging from strong disagreement (1) to neutrality (3) to strong agreement (5). Attitude scores are calculated by averaging or summing together responses to all the items.

Likert items measuring prejudice against whites.

_________________________________________________________________

Directions. Indicate degree of agreement by placing a digit in the blank preceeding an item where:

1 = strongly disagree      

2 = disagree

3 = neutral

4 = disagree

5 = strongly disagree

1. There is nothing lower than white trash.

2. It is usually a mistake to trust a white person.

3. If there is a heaven, it is hard to image that there are many white people up there.

4. The world might be a better place if there were fewe r white people.

________________________________________________________________

   (Source: Steckler, 1957)

A forth technique, called the semantic differential, was developed by Charles E. Osgood, in collaboration with several of his colleagues (Osgood, 1965; Osgood, Suci, & Tannenbaum, 1957), measured individuals' psychological definitions of various objects and concepts using sets of adjectives. As shown below each set consisted of two opposite adjectives--such as strong-weak, good-bad, clean-dirty--and subjects rated the object along the line between the two anchoring adjectives. According to Osgood, the semantic differential measures three aspects of any object: evaluation (good-bad, valuable- worthless), potency (strong-weak, heavy-light), and activity (fast-slow, excitable-calm). Many researchers feel that the evaluative factor corresponds most closely to the affective component of an attitude.

Semantic differential measures of men.

Please complete this questionnaire by placing an X on the line

between the dots.

                            Men Are:                       

            1.  beautiful  :__.__.__.__.__:  ugly    

            2.     strong  :__.__.__.__.__:  weak   

            3.     active  :__.__.__.__.__:  passive

            4.       good  :__.__.__.__.__:  bad

            5.      sharp  :__.__.__.__.__:  dull

            6.       fast  :__.__.__.__.__:  slow

            7.   valuable  :__.__.__.__.__:  worthless

            8. successful  :__.__.__.__.__:  unsuccessful

            9.  excitable  :__.__.__.__.__:  calm

(Items 1,4, and 7 measure the evaluative component; Items 2, 5, and 8 measure the potency component; items 3, 6, and 9 measure the activity component).

Multifactor scales. If researchers believe that an attitude is actually made up of a number of interrelated attitudes (or factors), then they can measure each factor separately with a multifactor scale. For example, one team of investigators (Woodmansee & Cook, 1967) proposed that white Americans' attitudes towards blacks were based on a number of factors, and that scales single factor scales overlooked this differentiation. In consequence, they developed the Multifactor Racial Attitude Inventory to measure 11 different factors underlying prejudicial attitudes. A sampling of these factors, and representative items, are shown below.

Sample items from the Multifactor Racial Attitude Inventory.

________________________________________________________________

_______________________________________________________________

(Source:  Woodmansee & Cook, 1967).

Self-report methods possess both weaknesses and strengths. They depend very much on knowing what questions to ask people. The observer may be able to incorporate previously unnoticed variables into the research scheme once the project is launched. The self- report researcher, in contrast, is likely either to fail to notice unexpected variables or to experience great difficulty in adapting the survey to include them. A maze of technical questions also confronts researchers designing questionnaires. If respondents don't answer the questions in consistently--if, for example, Jos indicates he dislikes Derlegians on Monday but on Tuesday says he likes them, then responses will be unreliable. Also, if questions aren't worded properly, the instrument will lack validity since respondents may misinterpret what is being asked. Validity is also a problem if group members aren't willing to disclose their personal attitudes, feelings, and perceptions or are unaware of these internal processes (Dawes & Smith, 1985).

Despite these limitations, self-report methods are indispensable research tools. Whereas observers may have difficulty determining individuals' private thoughts and emotions, survey researchers ask questions designed to tap specific areas. In addition, by asking the right questions of the right people, they can zero in on the specific problem of interest. They can ask questions that directly measure reports of feelings about others and the group, and the responses to such questions are relatively easy to interpret. Because questionnaire responses can be objec- tively coded, very specific conclusions can be drawn from questionnaire and interview responses. Thus, self-report methods are of value for two basic reasons: they can be used to tap variables directly that may be difficult to assess otherwise, and they can yield very specific conclusions about the relationships among variables.

Observation

Social anthropologist Erving Goffman tells of a visitor to the Shetland Islands dining with a local family. When the well- mannered visitor politely compliments his hostess on her fine soup, she nods in appreciation. But because he eats the soup very slowly and with great deliberation, the hostess decides that her guest really didn't like the soup at all (Goffman, 1959).

Behaviors sometimes provide more information than words alone, so researchers often prefer to record the actions undertaken by people in social settings. Observation requires watching and recording social behavior with subjects' knowledge (overt observation), from a hidden vantage point (covert observation), or as a member of the interaction itself (participant observation). When researchers count the number of times a person smiles when interviewing a job candidate (Deutsch, 1990), watch pedestrians jaywalking on a city street (Mullen, Cooper, & Driskell, 1990), or pretend to be trainees in order to discover the influence tactics taught to door-to-door salespeople (Cialdini, 1988) they are using observational methods.

William Foote Whyte (1943) relied on observational measures in his classic study of Italian-American gangs in the heart of Boston. He moved to the district, which he gives the fictitious name Cornerville, lived for a time with an Italian family, and joined the Nortons, a group of young men who gathered at a particular corner on Norton Street. He also participated in a more social club known as the Italian Community Club. Whyte observed and recorded these groups for 3« years, gradually developing a detailed portrait of this community and its groups. In particular, Whyte's study underscored a strong link between the individual member and the group. If a young man belonged to a corner gang such as the Nortons, his life was dramatically influenced by this group; he became a "corner boy" first, an individual second. Doc, the leader of the Nortons, pointed out that a corner boy would be lost without his gang.

They come home from work, hang on the corner, go up to eat, back on the corner, up to a show, and they come back to hang on the corner. If they're not on the corner, it's likely the boys there will know where you can find them. Most of them stick to one corner. It's only rarely that a fellow will change his corner [Whyte, 1943, p. 256].

Whyte's study also illustrates some key features of observational measures. Whyte, unlike some observers, actually joined the groups he studied. He also revealed his identity to the group: the group members knew he was a student at Harvard writing a book about the community. Whyte, at least initially, also decided to just observe as much of Cornerville life each day as he could, without narrowing his investigation by focusing on any specific types of behavior. These decisions, as noted below, all had a great impact on his final study and its conclusions.

Participant observation. Whyte decided to study the Boston slum groups from within; he actually joined these groups. This technique of group observation is called participant observation, which can be defined as "a process in which the observer's presence in a social situation is maintained for the purpose of scientific investigation. The observer is in a face-to-face relationship with the observed, and, by participating with them in their natural life setting, he gathers data" (Schwartz & Schwartz, 1955, p. 344; see also Schwartz & Jacobs, 1979).

Whyte, as a participant observer, gained access to information that would have been hidden from an external observer. His techniques also gave him a very detailed understanding of the gang. Rather than portraying the group as a static entity, he described groups as they are: dynamic interpersonal systems that evolve over time. He was also able to clarify formerly vague concepts such as leadership and power by tying these terms to directly observable, unambiguously described group events. Unfortunately, his presence in the group may also have changed the group itself. He went bowling with the Nortons, gambled with them, and even lent money to some of the members. His presence in the group undoubtedly modified its structure, and therefore the group he describes is not a typical corner gang, but, rather, a corner gang with a researcher in it (see Figure 2-1).

Overt and covert observation. Whyte was an overt observer of the group; the Nortons knew that he was a social scientist and that he was recording their behavior. Such openness meant that he did not have to mislead the group in any way, but by revealing his purpose he may have indirectly influenced the gang's behavior. As one corner boy once remarked, "You've slowed me down plenty since you've been down here. Now, when I do something, I have to think what Bill Whyte would want to know about it and how I can explain it. Before I used to do things by instinct" (p. 301).

The tendency for individuals act differently when they know they are being observed is known as the Hawthorne effect, after research conducted by Elton Mayo and his associates at the Hawthorne Plant of the Western Electric Company (Landsberger, 1958; Mayo, 1945; Roethlisberger & Dickson, 1939). Mayo and his colleagues wanted to understand productivity in the workplace, so they systematically varied a number of features while measuring the workers' output. For example, they moved one group of women to a separate room, where their performance could be carefully monitored. Next, they introduced factors into the situation that they thought would hurt performance, such as reduced lighting and fewer rest periods, and factors that he thought would help performance, such as brighter lighting and more rest periods. Surprisingly, all innovations led to improved worker output. Mayo, to his credit, recognized that the physical features he had manipulated were not as important as the social factors present in the work group. Apparently, the group members felt that the company was taking a special interest in them, and so they re- sponded by working particularly hard. (Varying views on these studies are presented by Bramel & Friend, 1981 and Franke, 1979; Franke & Kaul, 1978.)

Mayo's work underscored the importance of considering interpersonal factors in the work setting, but at the same time it warned investigators of a serious research problem: group members act differently when they believe they are being observed by social scientists interested in their behavior (Barnard, 1938; McGregor, 1960). Because the Hawthorne effect can limit the generalizability of research findings, some researchers use covert observation; the observer records the group's activities without the subject's knowledge. Such methods are commendable methodologically, but researchers face ethical issues if their observations invade the privacy of the people they are watching. Direct, nondeceptive methods avoid these moral quagmires but at the price of less accurate information (Humphreys, 1975; Sieber, 1992; Steininger, Newell, & Gracia, 1984).

Structuring observations. Whyte decided to observe the Nortons each day, but he did not organize his observations initially. He simply took extensive notes, which he later integrated to form an overall picture of the group. To some, such an open approach to observation is needed so that final conclusions are not biased by the researchers' preconceptions about the groups (Barton & Lazarfeld, 1969; Glaser & Strauss, 1967; Schwartz & Jacobs, 1979).

Others, however, argue that such openness should be avoided, because it puts too much trust in the observational powers of the researchers: they may let initial, though implicit, expectations shape their records (Mitroff & Kilmann, 1978; Weick, 1985). Albert H. Hastorf and Hadley Cantril (1954) demonstrated just such a perceptual bias in their classic "They Saw a Game" experiment. They asked college students to watch a film of two groups--two teams playing a football game. They selected a game between Dartmouth and Princeton that featured extremely rough play and many penalties against both teams. The Princeton quarterback, an all- American, left the game in the second quarter with a broken nose and a mild concussion. In the third quarter the Dartmouth quarterback's leg was broken when he was tackled in the backfield. Hastorf and Cantril asked Dartmouth and Princeton students to record the number and severity of the infractions that had been committed by the two teams.

As Hastorf and Cantril expected, some of their student observers weren't very objective. Dartmouth students saw the Princeton Tigers commit about the same number of infractions as the Dartmouth players. Princeton students, however, disagreed with the Dartmouth observers; they saw the Dartmouth team commit more than twice as many infractions as the Princeton team. Apparently, the Princeton observers' preference for their own team distorted their perceptions of the group interaction.

Structured observational measures offer one possible solution to the problem of objectivity in observations by helping observers categorize group behavior. Like biologists who classify living organisms under such categories as phylum, subphylum, class, and order or psychologists who classify people into various personality groupings or types, group researchers want to be able to classify each group behavior into an objectively definable category. They achieve this goal by first deciding which behaviors in the group are of interest and which are not. Next, they set up the categories to be used in the coding system. The researchers then note the occurrence and frequency of these targeted behaviors (Weick, 1985; Whitley, 1996).

Researchers can now choose from a wide assortment of structured observational systems; indeed, one researcher has actually developed a structured coding system for classifying types of coding systems (Trujillo, 1986). Robert Freed Bales's method has proven particularly useful (Bales, 1950, 1970, 1980). Bales's called his first coding system Interaction Process Analysis (IPA). Researchers who used IPA could classify each bit of behavior per- formed by a group member into 1 of 12 categories. Six of these categories (1-3 and 10-12) pertained socioemotional activities that sustained or weakened interpersonal relationships within the group. Complimenting another person would be an example of a positive socioemotional behavior, whereas insulting a group member would reflect negative socioemotional behavior. The other six categories pertain to task activity, or behavior that focuses on the problem the group is trying to solve. Giving and asking for information, opinions, and suggestions related to the problem the group faces are all examples of task-oriented activity.

Interaction Process Analysis system.

(Source: Bales, 1970)

To use the IPA, observers must learn to identify the 12 types of behavior defined by Bales. They must practice listening to a group discussion, breaking the verbal content down into the smallest meaningful units that can be identified, and then classifying these units as to category (Bales, 1950, p. 7). Thus, considerable training is involved in this form of observation. As the group members interact, observers record on a profile form containing the categories listed in the table who spoke to whom (for example, Sophia to Barbara) and the type of statement made (say, a statement that shows solidarity or seems friendly). If Sophia, for example, begins the group discussion by asking "Should we introduce ourselves?" and Barbara answers "Yes," the observers write 1-2 beside Category 8 (Sophia asks for opinion) and 2-1 beside Category 5 (Barbara gives opinion to Sophia). If later in the interaction Stephan angrily tells the entire group "This group is a boring waste of time," the coders write Stephan-0 beside Category 12 (Stephan seems unfriendly to entire group).

If observers are well-trained, a structured coding systems such as the IPA will yield data that is both reliable and valid. Reliability, in a measure, depends on its consistency: across time, components, and raters. If raters, when they hear the statement "This group is a boring waste of time," always classify it as an example of Category 12, then the measure is reliable. The measure also has interrater reliability if different raters, working independently, all think that the phrase belongs in Category 12. Validity describes the extent to which the technique measures what it is supposed to measure. To be valid, the IPA must actually measure the amount of task-oriented and socioemotional activity in the group. If the observers aren't correct in their coding or the categories don't accurately reflect these two aspects of a group, then the scores won't be valid.

The IPA is useful because it records the number of times a particular type of behavior has occurred and makes possible comparison across categories, group members, and even different groups. Bales has also improved both the reliability and validity of his coding system over the years. As the table indicates, Bales revised several of the categories in 1970 to increase their usefulness during observation, and even more recently he proposed a further elaboration of the entire system. This newest version is called SYMLOG, which stands for System of Multiple Level Observation of Groups (Bales, 1980, 1988). It is based on three structural dimensions assumed to underly differences among individuals in groups: dominance versus submission (or status), friendliness versus unfriendliness (or attraction), and instrumental control versus emotional expressiveness (or role orientation). With SYMLOG, observers must rate each action as an instance of 1 of 26 different categories of behavior. The categories, expressed in their most general form, follow.

1. active, dominant, talks a lot

2. extraverted, outgoing, positive

3. purposeful, democratic task-leader

4. assertive, business-like, manager

5. authority, controlling, critical

6. domineering, tough-minded, powerful

7. provocative, egocentric, showed-off

8. joked around, expressive, dramatic

9. entertaining, sociable, smiled, warm

10. friendly, egalitarian

11. worked cooperatively with the others

12. analytical, task-oriented, problem-focused

13. legalistic, wanted to be right

14. unfriendly, negativistic

15. irritable, cynical, didn't cooperate

16. showed feelings and emotions

17. affectionate, likeable, fun to be with

18. looked up to others, appreciative

19. gentle, willing to accept responsibility

20. obedient, worked submissively

21. self-punishing, worked too hard

22. depressed, sad, resentful, rejecting

23. alienated, quit, withdrawn

24. afraid to try, doubted own ability

25. quitely happy just to be in group

26. passive, introverted, said little

Given the greater reliability and validity of structured observations, why did Whyte take an unstructured approach? Whyte was more interested in gaining an understanding of the entire community and its citizenry, so a structured coding system narrow focus on specific behaviors would have yielded an unduly narrow analysis. Also, at the time he conducted his study, very little research had been done on community groups, so he did not know which behaviors he should scrutinize if he wanted to understand the group. Whyte was also personally unfamiliar with this area of Boston, and so he needed time to learn to "take for granted the same things that my Cornerville friends take for granted" (p. 321). He therefore chose to immerse himself in field work by taking extensive notes and waiting for some overall guiding theme to make itself evident. Thus, if the research is more exploratory, designed to develop theory first and validate hypotheses second, then an unstructured observational approach is appropriate. If the researcher is testing a specific hypothesis that can be tested by measuring specific aspects of a group, then the rigor and objectivity of a structured approach seems preferable.

Structured observation is a specialized form of observation; it guides observers' perceptions by identifying classes of target behaviors that are of particular interest to the researcher (Weick, 1985). A recent study of boy-meets-girl encounters nicely illustrates this procedure. College-student subjects waiting for an experimenter were secretly videotaped. These encounters were then analyzed in detail, with researchers coding such behaviors as body orientation, gaze, mutual gaze, hand gestures, smiles, laughter, and amount of time spent talking. The researchers discovered that these behaviors varied systematically depending on how shy the men were and how attractive the women were. Shy men were quieter and more likely to avoid eye contact; attractive women talked more, for longer periods of time, and made more eye contact with their partners. Note, too, that the researchers protected the privacy of the subjects by offering to erase the videotape before it was studied or reviewed in any way (Garcia et al., 1991).

Behavioral Measures

In rare cases behaviors can be measured through the use of mechanical recording devices, such as response counters, timers, and computers. Researchers assess aggression by arranging for pairs of subjects to give each other mild electric shocks (Baron, 1972) and productivity by counting the number of nonredundant ideas generated by brainstorming groups (Diehl & Stroebe, 1991). Social psychologists also use computers in their research when they need to precisely measure reaction time: the amount of time it takes a person to respond to stimulus information under controlled conditions. Reaction time is an indirect way to discover how people process information. If people take longer to respond to a question we can assume that their memory system to work harder to locate the answer.

Physiological Measures

We are not robots. Changes in our overt actions are matched by changes at the physiological level. When we see other people in distress and feel sympathy for them in their plight, our heart rates slow and blood pressure drops (Eisenberg et al., 1990). When we compete with others, our muscles relax and contract in distinctive ways, our body temperature and heart rate increase, and we tend to sweat more (Blascovich, Nash, & Ginburg, 1978; Lanzetta & Englis, 1989; Van Egeren, 1979). When we find ourselves in stressful situations our blood pressure and heart rate increase (Contrada, 1989), and merely watching videotaped scenes of violence and aggression are enough to trigger changes in heart rate (Bushman & Geen, 1990). Social-psychophysiological measures identify physiological correlates of social behavior, including the neurological, muscular, and respiratory systems that operate as we deal with other people.

Physiological systems can be difficult to measure for technical reasons, but many social psychophysiologists feel that the results justify the effort (Blascovich & Kelsey, 1990). Physiological changes can, for instance, be important outcomes of social processes; when we are fearful, heart rate increases, and when we are relaxed, it slows. Physiological changes can also be markers for hard-to-detect phenomena; by measuring the muscle activity of the face using an electromyograph (EMG), we can tell when a person experiences anger or happiness (Cacioppo et al., 1988; Cacioppo & Tassinary, 1990). Social-psychophysiological measures are particularly useful when subjects' self-reports or overt actions might be distorted by social pressures. In studies of prejudice, respondents often deny that they are biased against other races of people. Yet, when their physiological responses are assessed, they show telltale signs of arousal and distress when touched by a member of another racial group (Rankin & Campbell, 1955).

Nonreactive Measures

Researchers also use measures that have little or no impact on the participants in the research. Such measures are so unobtrusive that they are hidden in the background of social settings. These nonreactive measures are most useful when subjects may act strangely if they are aware that their behavior is being investigated. Researchers have assessed compliance with a local law prohibiting the consumption of alcoholic beverages by sifting through residents' garbage cans (Webb et al., 1981) and obedience to signs prohibiting graffiti by counting the number of messages scrawled on the signs themselves (Pennebaker & Sanders, 1976; Sechrest & Belew, 1983).

Archival Analysis

Archival analysis is a special type of nonreactive measure that takes advantage of existing records. Often these records are in public archivesþthe reports and statistics that are routinely collected by government agencies. Other sources of archival data include old newspapers, personal documents, and the public speeches of politicians. In an application of this method to social psychology, researchers compared the number of times two theoretically important worksþLeon Festinger's 1957 book, A Theory of Cognitive Dissonance, and Fritz Heider's 1958 book, The Psychology of Interpersonal Relationshipsþwere cited by other social psychologists in scientific articles published from 1958 to 1988. They discovered that while references to Festinger's book have dwindled in the last 15 years, references to Heider's book on attribution and social cognition have remained numerous and constant (Bagby, Parker, & Bury, 1990).

Evaluating Measurement Methods

When social psychology was a fledgling discipline, many believed that it would never become a science because social processes such as impression formation, attitudes, love, conformity, and group behavior couldn't be measured. Over the years, however, researchers proved the pessimists wrong by developing the self-report, behavioral, physiological, and nonreactive measurement techniques.

No index of social processes is perfect. Self-report methods are well suited for studies of personal processes, such as thoughts, feelings, or attitudes. Unfortunately, these methods assume that people are aware of their intrapersonal processes and are willing to report them. If people can't gauge their own tendency to obey others, or if they don't want to admit it, researchers may want to turn to the study of behavioral processes. Observational techniques are excellent tools for studying human interaction, but if observers' attitudes and motivations influence their observations. the results may be biased. In this case, structured observations carried out by uninvolved, highly trained observers yield more objective information. Furthermore, when direct measures make subjects feel self-conscious, physiological and nonreactive measures may be the most appropriate approaches.

Other Notes in this Category

1. Attitudes
2. Attitudes and Behaviour
3. Attribution
4. Conformity
5. Crowds and Territoriality
6. Goffman - Symbolic Interactionist
7. Impression Formation
8. Inference Model
9. Intuition model
10. Is there a conformity personality?
11. Obedience
12. Persuasion
13. Prejudice
14. Self Concept
15. Self-attribution processes
16. Social Influence
17. Stereotypes and Stereotyping
18. The Primacy-Recency effect
19. Zimbardo

Didn't find this useful?

• Visit Coursework.Info for over 14,000 GCSE, A-Level and University Essays