2.2^2. Research Settings and Types of Research^37^50^,,^3416^3716%
Research Settings and Types of Research
Discuss common research settings and the three types of research that are used in psychology.

The collection of data is the fundamental means of testing hypotheses. In this section we investigate the major ways of gathering data about behavior and mental processes. We will find that three basic types of research are used in psychology: descriptive, correlational, and experimental.

The Scientific Method
(Your score will be reported to your instructor)
Page 38

One concept that is central to all of these research approaches is that of a variable, a term referring to anything that varies. In a person, variables might include height, weight, IQ, religious faith, and the extent to which the individual feels happy or unhappy, for example. We can consider anything that differs among people or changes within a person to be a variable. In general, all forms of scientific inquiry in psychology are interested in how variables relate to one another. The type of methods researchers use is typically guided by their conceptual understanding of the variables of interest.

Another fundamental element of any research is the setting where such research may take place. We begin our look at common approaches to psychological research by considering the settings available to researchers.

Research Settings

All three types of research that we will discuss shortly—descriptive, correlational, and experimental—can be carried out in different settings. In other words, the setting of the research does not determine the type of research it is. Common settings include the research laboratory and natural settings.

Classrooms, playgrounds, sports arenas, shopping malls, and other places that people live in and frequent serve as ideal settings for naturalistic observations.

Because psychological researchers often need to control certain factors that determine behavior but are not the focus of inquiry, much of their research is conducted in a laboratory, a controlled setting with many of the complex factors of the real world removed (Mitchell & Jolley, 2007).

Although laboratory research allows a great deal of control, doing research in the laboratory has some drawbacks. First, it is almost impossible to conduct research in the lab without the participants' knowing they are being studied. Second, the laboratory setting is unnatural and therefore can cause the participants to behave unnaturally. A third drawback of laboratory research is that people who are willing to go to a university laboratory may not fairly represent groups from diverse cultural backgrounds. Those who are unfamiliar with university settings and with the idea of “helping science” may be intimidated by the setting. Fourth, some aspects of the mind and behavior are difficult if not impossible to examine in the laboratory. Laboratory studies of certain types of stress may even be unethical.

Research can also take place in a natural setting. Naturalistic observation provides insight that researchers sometimes cannot achieve in the laboratory (Bronfenbrenner & Morris, 2006). Naturalistic observation is observing behavior in real-world settings. Psychologists conduct naturalistic observations at sporting events, day-care centers, work settings, shopping malls, and other places that people frequent. Suppose that you wanted to study the level of civility on your campus. Most likely, you would include some naturalistic observation of how people treat one another in such gathering places as the cafeteria and the library reading room.

Naturalistic Observation
(Your score will be reported to your instructor)
naturalistic observationObservation of behavior in real-world settings with no effort made to manipulate or control the situation.
Jane Goodall was a young woman when she made her first trip to the Gombe Research Center in the African country of Tanzania. Fascinated by chimpanzees, she dreamed about a career that would allow her to explore her hunches about the nature of chimpanzees. A specialist in animal behavior, she embarked on a career in the bush that involved long and solitary hours of careful, patient observation. Her observations spanned 30 years, years that included her marriage, the birth of her son, untold hardship, and inestimable pleasure. Due to her efforts, our understanding of chimpanzees in natural settings dramatically improved.

Naturalistic observation was used in one study that focused on conversations in a children's science museum (Crowley & others, 2001). The researchers' finding that parents were three times as likely to engage boys than girls in explanatory talk while visiting different exhibits suggests a gender bias that encourages boys more than girls in science (Figure 2.1). In another study, Mexican American parents who had completed high school used more explanations with their children when visiting a science museum than Mexican American parents who had not completed high school (Tenenbaum & others, 2002). Naturalistic observation allows the researcher access to a person's spontaneous behaviors; however, a key weakness of this method is the lack of control over the setting. For instance, imagine setting up one's research in a museum and having no one come by that day.

Parents' Explanations of Science to Sons and Daughters at a Science Museum
In a naturalistic observation study at a children's science museum, parents were three times more likely to explain science to boys than to girls (Crowley & others, 2001). The gender difference occurred regardless of whether the father, the mother, or both parents were with the child, although the gender difference was greatest for fathers' science explanations to sons and daughters.
Page 39
Descriptive Research

Some important psychological theories have grown out of descriptive research, which serves the purpose of observing and recording behavior. For example, a psychologist might observe the extent to which people are altruistic toward one another. By itself, descriptive research cannot prove what causes some phenomenon, but it can reveal important information about people's behaviors and attitudes. Descriptive research methods include observation, surveys and interviews, standardized tests, and case studies.


Imagine that you are interested in studying how children who are playing a game resolve conflicts that come up during the game. Thus, the data you are interested in concern conflict resolution. As a first step, you might go to a playground and simply observe what the children do—how often you see conflict resolution occur and how it unfolds. You would likely keep careful notes of what you observe.

This type of scientific observation requires an important set of skills. Unless we are trained observers and practice our skills regularly, we might not know what to look for, we might not remember what we saw, we might not realize that what we are looking for is changing from one moment to the next, and we might not communicate our observations effectively. Furthermore, it might be important to have more than one person do the observations as well, to develop a sense of how accurate your observations are. For observations to be effective, they must be systematic. We must have some idea of what we are looking for. We must know whom we are observing, when and where we will observe, and how we will make the observations. And we need to know in what form they will be recorded: in writing, by sound recording, or by video.

Surveys and Interviews

Sometimes the best and quickest way to get information about people is to ask them for it. One technique is to interview them directly. A related method that is especially useful when information from many people is needed is the survey, or questionnaire. A standard set of questions is used to obtain people's self-reported attitudes or beliefs about a particular topic. In a good survey, the questions are clear and unbiased, allowing respondents to answer unambiguously.

Surveys and interviews can probe into a wide range of topics, from religious beliefs to sexual habits to attitudes about gun control (Rosnow & Rosenthal, 2008). Surveys and interviews are conducted in person, by telephone, or (increasingly) over the Internet.

Some survey and interview questions are unstructured and open-ended, such as “Could you elaborate on your optimistic tendencies?” and “How fulfilling would you say your marriage is?” They allow for unique responses from each person surveyed. Other survey and interview questions are more structured and ask about quite specific things. For example, a structured survey or interview question might ask, “How many times have you talked with your partner about a personal problem in the past month: 0, 1–2, 3–5, 6–10, 11–30, every day?”

One problem with surveys and interviews is the tendency of participants to answer questions in a way that they think is socially acceptable or desirable rather than in a way that communicates what they truly think or feel (Nardi, 2006). For example, a person might exaggerate the amount of communication that goes on in a relationship in order to impress the interviewer.

One example of a survey conducted by the Gallup organization (1999) asked parents their beliefs about the most important problems facing schools. Forty-three percent cited drugs, 40 percent sex, 39 percent discipline in the classroom, 28 percent violence, and 25 percent social pressure among students to be popular. The survey was based on telephone interviews with a randomly selected sample of 338 U.S. parents. Recall the discussion of random sampling earlier in the chapter. When surveys are conducted on a national basis, as Gallup polls are, random sampling is considered to be an important aspect of the survey process.

Page 40
PSYCHOLOGY AND LIFE Who Is the Healthiest Person You Know?

One way to study psychological variables is through case studies, or case histories. Researchers often use case studies to understand individuals who suffer from psychological disorders. But they might similarly use the case history approach to plumb the reasons why people are psychologically healthy.

Imagine that you have been asked to do a case study of psychological wellness. Think of the psychologically healthiest and happiest person you know. Consider the following questions about your hypothetical study:

  • What makes this person a good example for a study of psychological health?

  • How would you gather data for your case study?

  • If you interviewed this person, what sorts of questions would you ask?

  • What might we learn about psychological health, broadly speaking, from such a study?

Standardized Tests

A standardized test requires people to answer a series of written or oral questions or sometimes both (Gregory, 2007). A standardized test has two distinct features: An individual's answers are tallied to yield a single score, or set of scores, that reflects something about that individual; and the individual's score is compared with the scores of a large group of similar people to determine how the individual responded relative to others (Gronlund, 2006). One widely used standardized test in psychology is the Stanford-Binet intelligence test, which we consider in Chapter 9.

standardized testA test that requires people to answer a series of written or oral questions or sometimes both.

Scores on standardized tests are often stated in percentiles. Suppose you scored in the 92nd percentile on the Scholastic Assessment Test (SAT). This score would mean that 92 percent of a large group of individuals who previously took the test received scores lower than yours.

The main advantage of standardized tests is that they provide information about individual differences among people. But one problem with standardized tests is that they do not always predict behavior in nontest situations. Another problem is that standardized tests are based on the belief that a person's behavior is consistent and stable, yet personality and intelligence—two primary targets of standardized testing—can vary with the situation. For example, a woman may perform poorly on a standardized intelligence test in an office setting but score much higher at home, where she is less anxious.

This criticism is especially relevant for members of minority groups, some of whom have been inaccurately classified as mentally retarded on the basis of their scores on intelligence tests (Hodapp & Dykens, 2006). In addition, cross-cultural psychologists caution that many psychological tests developed in Western cultures might not be appropriate in other cultures (Shiraev & Levy, 2007). People in other cultures may have had experiences that cause them to interpret and respond to questions much differently than the people on whom the test was standardized.

Standardized tests require individuals to answer a series of written or oral questions. The individual on the left is taking a standardized test of intelligence.
Case Studies

A case study, or case history, is an in-depth look at a single individual. Case studies are performed mainly by clinical psychologists when, for either practical or ethical reasons, the unique aspects of an individual's life cannot be duplicated and tested in other individuals (Dattilio, 2001). A case study provides information about one person's goals, hopes, fantasies, fears, traumatic experiences, family relationships, health, or anything else that helps the psychologist understand the person's mind and behavior. Sigmund Freud developed his theory of psychoanalysis based entirely on case studies of individuals suffering from psychological problems.

case studyAn in-depth look at a single individual; also known as a case history.
Page 41

Another example of a case study is the analysis of India's spiritual leader Mahatma Gandhi by psychodynamic theorist Erik Erikson (1969). Erikson studied Gandhi's life in great depth to discover insights into how his positive spiritual identity developed, especially during his youth. In putting together the pieces of Gandhi's identity development, Erikson described the contributions of culture, history, family, and various other factors that might affect the way other people develop an identity.

Case histories provide dramatic, detailed portrayals of people's lives, but we must be cautious when generalizing from this information. The subject of a case study is unique, with a genetic makeup and personal history that no one else shares. In addition, case studies involve judgments of unknown reliability. However, case studies may be useful in generating ideas that could then be tested in empirical investigations using larger samples and correlational or experimental designs. To get a taste of how researchers approach a case study, see Psychology and Life.

Mahatma Gandhi was the spiritual leader of India in the middle of the twentieth century. Erik Erikson conducted an extensive case study of his life to determine what contributed to his identity development.
Correlational Research
(Your score will be reported to your instructor)

Some psychological research relies on the systematic observation of variables within a sample of individuals. These studies are concerned with identifying the relationships between two or more variables in order to describe how these variables change together. This work is often called correlational research because of the statistical technique, referred to as correlation, that is typically used to analyze this type of data. The more strongly the two events are correlated (or related or associated), the more effectively we can predict one event from the other. The distinguishing feature of a correlational study is that the variables of interest are measured, not manipulated, by the researcher. That is, the researcher simply measures the variables of interest to see how they relate. No attempt is made to change the value of any of the variables.

correlational researchA research strategy that identifies the relationships between two or more variables in order to describe how these variables change together.

The degree of relationship between two variables is expressed as a numerical value called a correlational coefficient. Let's assume that we have data on the relationship between how many hours individuals spend volunteering for a variety of charities (the X variable) as well as the level of life satisfaction these people experience (the Y variable). For the sake of this example, let's assume these data produce a correlation coefficient (represented by the letter r) of +.70. Remember this number, as we will soon use it to illustrate what a correlation coefficient tells you about the relationship between two events or characteristics.

For the moment, however, you need to know only that the number tells you the strength of the relationship between the two variables. The rule is simple: The closer the number is to 1.00, the stronger the correlation; conversely, the closer the number is to .00, the weaker the correlation. Figure 2.2 offers guidelines for interpreting correlational numbers. But perhaps you are wondering about the significance of the plus sign in the correlation coefficient of +.70 that we have calculated in our classroom study.

Guidelines for Interpreting Correlational Numbers
The magnitude of a correlation tells us about the strength of the relationship between two variables.
Positive and Negative Correlations

The numerical value of a correlation coefficient always falls within the range from -1.00 to +1.00. The number of the correlation tells us about the strength of the relations, but the sign (+ or -) tells us about the direction of the relationship between the variables. So, negative numbers do not indicate a lower value than positive numbers. A correlation of -.65 is just as strong as a correlation of +.65. The plus or minus sign tells you nothing about the strength of the correlation. A correlation coefficient of -.87 is closer to -1.00 and thus indicates a stronger correlation than the coefficient of +.45 is to +1.00.

What the plus or minus sign does tell you is the direction of the relationship between the two variables. A positive correlation is a relationship in which the two factors vary in the same direction. Both factors tend to increase together, or both factors tend to decrease together. So, in the example above, the more time spent volunteering, the more satisfied people were with their lives. In addition, the positive correlation means that people who spent little time volunteering also showed lower life satisfaction. Either relationship represents a positive correlation. A negative correlation, in contrast, is a relationship in which increases in one variable are associated with decreases in the other. For instance, we might find that the number of hours spent watching TV is negatively correlated with life satisfaction. That means that the more TV a person watches, the lower his or her life satisfaction might be expected to be. Examples of scatter plots showing positive and negative correlations appear in Figure 2.3.

Scatter Plots Showing Positive and Negative Correlations
A positive correlation is a relationship in which two factors vary in the same direction, as shown in the two scatter plots on the left. A negative correlation is a relationship in which two factors vary in opposite directions, as shown in the two scatter plots on the right.
Page 42

An example of a correlational research on happiness is a set of studies conducted by psychologists Sonja Lyubomirsky and Lee Ross. These researchers were interested in how happy and unhappy people feel about decisions they have made in their lives. Imagine, for example, that you are going to buy a new computer. You browse through plenty of stores, talk to friends, and read product reviews in Consumer Reports magazine. Finally, you make your choice. How are you likely to feel about it, and how will you feel about the computers you almost bought?

Lyubomirsky and Ross (1999) conducted a series of studies investigating the relationship between being happy and evaluations of one's choices as well one's feelings about what he or she did not get. In one study, Lyubomirsky and Ross studied high school students who were applying to colleges. These students completed a measure of how generally happy they were and provided information about the colleges to which they were applying. They also rated how positively they felt about each school. Three months later, after acceptance and rejection letters had been sent out (and students' decisions about which school to attend had been made), the students once again rated the schools to which they had applied. How did knowing that they were accepted, rejected, going to a particular school, or not going influence their feelings?

With regard to the college they had chosen to attend, happy students tended to be very excited about their chosen school—their assessment of the school had become even more positive. In contrast, unhappy students did not show an increase in their feelings about the school they would be attending. Indeed, unhappy students tended to react somewhat negatively toward the school. With regard to schools that had accepted them but that they had declined to attend, happy students remained very positive about them. These schools, after all, had shown very good taste and judgment—the students' attendance there just was not meant to be.

Page 43

In contrast, unhappy students tended to denigrate schools that had accepted them but that they had chosen not to attend. The comedian Groucho Marx once quipped that he would never want to join a club that would have him as a member, and the unhappy students tended to reflect this outlook: Any school that would take them could not be that great.

In a second study, conducted in the laboratory, happy and unhappy participants were presented with 10 fancy desserts to evaluate before and after they were told which one they would get to eat (Lyubomirsky & Ross, 1999). Participants were read descriptions of 10 different delicious desserts—cheesecake, lemon poppy seed cake, chocolate-on-chocolate cake, carrot cake, and so on. They were asked to rank the desserts in terms of how much they would want to try them. Participants were always told that they would be receiving their second choice. Before getting their dessert, however, participants were asked to rate the choices one more time—at this point they were presented with pictures of the desserts.

Once again, happy people showed a tendency to increase their liking of the dessert they were told they would be receiving, while unhappy people actually slightly decreased their liking for the dessert when they realized they would be getting it. In addition, when rating a dessert they would not be receiving, the unhappy people tended to derogate (that is, to belittle) the dessert; that slice of chocolate-on-chocolate cake suddenly looked much less appetizing when they knew it was not going to be theirs. These findings provide insight into the different thinking styles of happy and unhappy people. Compared to happy people, unhappy individuals appear to feel compelled to derogate alternatives they did not choose or could not have. Happy people, in contrast, appear to live in a world full of good things, and they tend to celebrate what they have received without feeling the need to boost its value by devaluing alternatives.

These studies are correlational in nature because in both studies happiness was measured, not manipulated—it was a variable that the participants brought with them into the studies. Lyubomirsky and Ross measured two variables: happiness (operationalized here by ratings on a happiness scale) and postdecisional judgments of alternatives that were either available or not (the ratings made of colleges and desserts before and after choices were made). They found that, compared to unhappy people, happy people were more likely to celebrate what they were getting (whether it was a college to attend or a dessert to eat). We might state this research result as, “Happiness was positively correlated with positivity about one's choices.” As happiness increased, so did positive ratings of chosen schools and desserts. In addition, happiness was negatively correlated with derogating alternatives that were not chosen. As happiness increased, the tendency to view unchosen alternatives negatively decreased.

This example clarifies some key points about systematic observation. First, the setting of a study does not determine its methodology. Correlational studies can take place in a classroom, out in the world, or in a laboratory. Also, the methods for data analyses do not determine the design of a study. Lyubomirsky and Ross did not use correlation coefficients to analyze their data. Even though they are typically referred to as correlational studies, such investigations need not rely on correlation coefficients. Once again, the key defining feature of this type of research is that the variables are simply measured.

Correlation and Causation

In trying to make sense of the world, people often make a big mistake about correlation. Look at the terms in bold type in the following newspaper headlines:

  • Researchers Link Coffee Consumption to Cancer of Pancreas

  • Scientists Find Connection Between Ear Hair and Heart Attacks

  • Psychologists Discover Relationship Between Marital Status and Health

Reading these headlines, the general public would tend to jump to the conclusion that coffee causes cancer, ear hair causes heart attacks, and so on. But all of the words in bold type are synonymous only with correlation, not with causality. Correlation does not equal causation. Remember, correlation means only that two variables change together. Being able to predict one event based on the occurrence of another event does not necessarily tell us anything about the cause of either event (Howell, 2008). Sometimes an extraneous variable that has not been measured accounts for the relationship between two others. This circumstance is referred to as the third variable problem.

third variable problemThe situation where an extraneous variable that has not been measured accounts for the relationship between two others.
Page 44

To understand the third variable problem, consider the following example. A researcher measures two variables: the number of ice cream cones sold in a town and the number of violent crimes that occur in that town throughout the year. She finds that ice cream cone sales and violent crimes are positively correlated, to the magnitude of +.50. This high positive correlation would indicate that as ice cream sales increase, so do violent crimes. Would it be reasonable for the local paper to run the headline “Ice Cream Consumption Leads to Violence”? Should protesters set themselves up outside the local Frosty Freeze to stop the madness? Probably not. Perhaps you have already thought of the third variable that might explain this correlation—heat. Indeed, when it is hot outside, people are more likely both to purchase ice cream and to engage in aggressive acts (Anderson & Bushman, 2002). These “third variables” are also called confounds.

Given the potential problems with extraneous third variables, why do researchers conduct correlational studies? There are several very good reasons. One reason is that some important questions can be investigated only by using correlational designs. Such questions may involve variables that cannot be manipulated, such as biological sex, personality traits, genetic factors, and ethnic background. Another reason why researchers conduct correlational studies is that sometimes the variables of interest are real-world events that influence people's lives, such as the effects of the September 11 attack of the World Trade Center on New York City residents. Correlational research is also valuable in cases where it would not be ethical to carry out an experiment because of the dangers it poses, such as one in which expectant mothers are directed to smoke varying numbers of cigarettes so that the researcher can see how cigarette smoke affects birth weight and the fetal activity level.

Correlational studies are useful, too, when the issue under investigation is post hoc (after the fact) or historical, such as research into the childhood backgrounds of people who are particularly successful. And correlational research is used when researchers are interested in everyday experience, which is difficult to study by bringing people into the artificial setting of a laboratory. For example, correlational researchers have begun to use daily diary methodologies, known as the experience sampling method (ESM), to study people in their natural settings. These studies involve having people document their daily experiences in a diary a few times a day, or complete measures of their mood and behavior whenever they are beeped by an electronic organizer. One recent daily diary study examined the experience of the meaning in life on a daily basis. In this study, student volunteers rated their mood, activities, thoughts, and their sense of meaning in their life twice daily for a week. The data showed that the strongest predictor of a day being felt to be meaningful was the amount of positive mood the person experienced that day (King & others, 2006).

One way that correlational researchers can confront the problem of third variables is to include these variables in designs that adopt a multivariate approach—a method that involves more than just the two main variables of interest. If a variable is measured, it can be controlled for, if not experimentally then statistically. In this way, for example, we can show that the number of cigarettes a person smokes does predict the likelihood of developing lung cancer, controlling for such factors as alcohol consumption, diet, body weight, family background, exercise, and so on. Thus, although correlation cannot be assumed to imply a causal relationship, correlational studies can prove very useful in pinning down potentially causal relationships by employing multivariate approaches to problems that are difficult to study through experiments.

An interesting research question that has been addressed in this way is, Do happy people live longer? In one study, 2,000 older Mexican Americans were interviewed twice over the course of 2 years (Ostir & others, 2000). In the first assessment, participants completed measures of happiness but also reported about potential third variables such as diet, physical health, smoking, marital status, and distress. Two years later, the researchers contacted the participants again to see who was still alive. Results showed that, controlling for these many potential third variables, happiness predicted who was alive 2 years later.

Page 45

Another way that correlational researchers can approach the issue of causation is to employ a special kind of systematic observation called a longitudinal design. Longitudinal research involves obtaining measures of the variables of interest in multiple waves over time. Longitudinal research can address the issue of causation because we can assume that if variable X causes changes in variable Y, X ought to, at least, precede Y over time.

longitudinal designA special kind of systematic observation that involves obtaining measures of the variables of interest in multiple waves over time.

One intriguing longitudinal study is the Nun Study, conducted by David Snowdon and his colleagues (Riley & others, 2005; Snowdon, 2003; Tyas & others, 2007). The study began in 1986 and has followed a sample of 678 School Sisters of Notre Dame ever since. The nuns ranged in age from 75 to 103 when the study began. These women complete a variety of psychological and physical measures annually. This sample is, of course, unique in many respects. However, some characteristics render the participants an excellent group for correlational research. For one thing, many potential extraneous third variables are relatively identical for all the women in the group. Their gender, living conditions, diet, activity levels, marital status, and religious participation are essentially held constant, providing little chance that differences in these variables can explain results. (If a variable does not change, it cannot change with or correlate with anything else.)

Researchers recently examined the question of the relation between happiness and longevity using this rich dataset. All of the nuns had been asked to write a spiritual autobiography when they entered the convent (for some, as many as 80 years before). Deborah Danner and her colleagues were given access to these documents and used them as indicators of happiness earlier in life, by counting the number of positive emotions expressed in the autobiographies (Danner, Snowdon, & Friesen, 2001). (Note that here we have yet another operational definition of happiness.) Higher levels of positive emotion expressed in autobiographies written at an average age of 22 were associated with a 2.5-fold difference in risk of mortality when the nuns were in their 80s and 90s. That is, women who included positive emotion in these autobiographies when they were in their early 20s were two-and-a-half times more likely to survive some 60 years later.

The use of multivariate approaches and longitudinal designs are ways that correlational researchers may attempt to indicate causal relations among variables. These are the kinds of studies that, along with experimental research using animal models, have allowed researchers to conclude that cigarette smoking causes cancer. Still, it is important to keep in mind that even the brightest scientist may not think of all of the potential third variables that might explain her results. Throughout this book you will read about numerous correlational research studies. Keep in mind how easy it is to assume causality when two events or characteristics are merely correlated. Think about those innocent ice cream cones, and critically evaluate conclusions that may be drawn from simple observation.

Experimental Research
Designing an Experiment: Dependent and Independent Variables
(Your score will be reported to your instructor)

If two variables are correlated, there might be a causal relationship between them; but if there is, we cannot be sure which way the causal arrow ought to point. Does X cause Y, or does Y cause X? Recent research on meaning in life provides a case in point.

Experiencing one's life as meaningful has long been assumed to be an important aspect of psychological well-being (Frankl, 1984; Steger & Frazier, 2005). Because measures of meaning in life and well-being have been shown to correlate positively (that is, the more meaning in life you have, the happier you are), the assumption has been that meaning in life caused the greater happiness. But because the studies involved in exploring this relationship have been correlational in nature, the causal pathway might well run in the other direction: Happiness might make people feel that their lives are more meaningful. A series of laboratory experiments has shown this very outcome. Laura King and colleagues (2006) have demonstrated that putting people in a good mood—by having them imagine themselves being recognized as a hero for helping a lost child find his parents—caused them to rate their lives as more meaningful than individuals who were told to imagine a neutral experience.

Page 46

To clarify the direction of causality, then, psychologists who are interested in determining the causal relationships that might exist between variables must turn to experimental methods (McBurney & White, 2007). An experiment is a carefully regulated procedure in which one or more variables believed to influence the behavior being studied are manipulated while all other variables are held constant.

experimentA carefully regulated procedure in which one or more variables believed to influence the behavior being studied are manipulated while all other variables are held constant.

If the behavior under study changes when a variable is manipulated, we say that the manipulated variable has caused the behavior to change. In other words, the experiment has demonstrated cause and effect. In the example above, positive mood was the cause, and meaning in life was the effect. This notion that experiments can demonstrate causation is based on the idea that if participants are randomly assigned to groups, the only systematic difference between them must be the manipulated variable. Random assignment means that researchers assign participants to groups by chance. This technique reduces the likelihood that the experiment's results will be due to any preexisting differences between groups (Martin, 2004). In the case of the study of meaning in life by King and others, because of random assignment we can assume that the groups (positive versus neutral mood) did not differ in meaning in life from the outset.

random assignmentThe assignment of participants to research groups by chance.
Independent and Dependent Variables

Experiments have two types of variables: independent and dependent. An independent variable is a manipulated experimental factor. It is a potential cause. The label “independent” is used because this variable can be manipulated independently of other factors to determine its effect. Researchers have a vast array of options open to them in selecting independent variables, and one experiment may include several independent variables. In the study of positive mood and meaning in life, the independent variable is mood (positive versus neutral).

independent variableThe manipulated experimental factor in an experiment.

A dependent variable is a factor that can change in an experiment in response to changes in the independent variable. As researchers manipulate the independent variable, they measure the dependent variable for any resulting effect. In the study of mood and meaning in life, meaning in life is the dependent variable.

dependent variableA factor that can change in an experiment in response to changes in the independent variable.
Experimental and Control Groups

Experiments can involve one or more experimental groups and one or more control groups. An experimental group is a group whose experience is manipulated. A control group is as much like the experimental group as possible and is treated in every way like the experimental group except for the manipulated factor. The control group thus serves as a baseline against which the effects of the manipulated condition can be compared.

control groupA comparison group that is as much like the experimental group as possible and is treated in every way like the experimental group except for the manipulated factor.
experimental groupA group in the research study whose experience is manipulated.
Some Cautions About Experimental Research

Validity refers to the soundness of the conclusions we draw from an experiment. Two types of validity matter to experimental designs. The first is ecological validity, which refers to the extent to which an experimental design is representative of the real-world issues it is supposed to address. That is, do the experimental methods and the results generalize to the real world?

ecological validityThe extent to which an experimental design is representative of the real-world issues it is supposed to address.
validityThe soundness of the conclusions we draw from an experiment.

Imagine that a researcher is interested in the influence of mood on creative problem solving. She randomly assigns individuals to listen to happy music (a positive mood induction) or sad music (a negative mood induction). She then gives all participants a chance to be creative by listing all of the uses they can think of for a cardboard box. Counting up the number of uses that people list, she finds that those in the happy-mood condition have generated more uses for the box. This finding might indicate that happiness is related to creativity. Considering the ecological validity of this study, we might ask the questions, How similar is the happy mood of participants in this study to the kinds of happy moods people experience in real life? and How much is listing the uses of a cardboard box really a sign of creativity? In other words, we ask, Do these methods do a good job of reflecting the real-world processes they are supposed to represent?

The second type of validity is internal validity, which refers to the extent to which changes in the dependent variable are due to the manipulation of the independent variable. In this case, we want to know if the experimental methods are free from biases and logical errors that may render the results suspect. Although experimental research is a powerful tool, it requires safeguards (Leary, 2008). Expectations and biases can, and sometimes do, tarnish results (Rosnow & Rosenthal, 2008).

internal validityThe extent to which changes in the dependent variable are due to the manipulation of the independent variable.
Page 47
Experimenter Bias

Experimenters may subtly (and often unknowingly) influence their research participants. Experimenter bias occurs when the experimenter's expectations influence the outcome of the research.

experimenter biasThe influence of the experimenter's own expectations on the outcome of the research.

In a classic study, Robert Rosenthal (1966) turned college students into experimenters. He randomly assigned the participants rats from the same litter. However, half of the students were told that their rats were “maze bright,” whereas the other half were told that their rats were “maze dull.” The students then conducted experiments to test their rats' ability to navigate mazes. The results were stunning. The so-called maze-bright rats were more successful than the maze-dull rats at running the mazes. The only explanation for the results is that the college students' expectations affected the performance of the rats. In subsequent studies, researchers have demonstrated that experimenters' expectations influence not only rodent behavior but human behavior as well (Rosenthal, 1994).

Research Participant Bias and the Placebo Effect

Like the experimenters, research participants may have expectations about what they are supposed to do and how they should behave, and these expectations may affect the results of experiments (L. Christensen, 2007). Research participant bias occurs when the behavior of research participants during the experiment is influenced by how they think they are supposed to behave.

research participant biasThe influence of research participants' expectations on their behavior within an experiment.

For example, in one study, the researchers first assessed participants' sensitivity to pain (Levine, Gordon, & Fields, 1979). Then they gave the participants an injection of a painkiller, or so the participants thought. Actually, they received a placebo a harmless, inert substance that has no specific physiological effect. (A placebo can be given to participants instead of the presumed active agent, such as a drug, to determine if the placebo produces the effects thought to characterize the active agent.) Subsequently, when the experimenter administered painful stimuli, the participants perceived less pain than they had in the earlier assessment of their sensitivity to pain. This experiment demonstrated a placebo effect, which occurs when participants' expectations, rather than the experimental treatment, produce an experimental outcome.

placebo effectThe situation where participants' expectations, rather than the experimental treatment, produce an experimental outcome.
placeboA harmless, inert substance that may be given to participants instead of a presumed active agent, such as a drug, and that has no specific physiological effect.

U.S. television viewers are often exposed to advertisements for prescription drugs. These ads typically include a voice-over that describes the potential side effects. You may have heard the statement “Some individuals taking this drug complain of headaches or stomach discomfort, but these effects were no different from those experienced by people receiving the placebo or sugar pill.” Experimenters use placebos to ensure that the effects of a medication are not simply due to expectations. Placebo effects can be surprisingly strong. Research has shown that a substantial part of the treatment effects for antidepressants, for example, may come out of the beliefs of the doctors and patients who use them (Kirsch & Sapirstein, 1999).

Advertisements for prescription drugs usually describe not only the side effects on people taking the actual drug but also the effects experienced by individuals receiving a placebo.

Another way to ensure that neither the experimenter's nor the participants' expectations affect the outcome is to design a double-blind experiment. In this design, neither the experimenter nor the participants are aware of which participants are in the experimental group and which are in the control group until the results are calculated. A study of drug treatment for social phobia was conducted in a double-blind manner (Van Ameringen & others, 2001). Both the experimenter who administered the drug and the participants were kept in the dark about which individuals were receiving the drug and which were receiving a placebo that looked like the drug. This setup ensured that the experimenter could not, for example, make subtle gestures signaling who was receiving the drug and who was not. A double-blind study allows researchers to tease apart the actual effects of the independent variable from the possible effects of the experimenter's and the participants' expectations about it.

double-blind experimentAn experiment that is conducted so that neither the experimenter nor the participants are aware of which participants are in the experimental group and which are in the control group until after the results are calculated.
Page 48
INTERSECTION Anthropological Theory and Social Psychology: Can Reminders of Death Influence Political Allegiance?

At the beginning of this chapter, we considered a variety of life experiences that seem difficult to study through empirical research. We have found that even for something as potentially indescribable as happiness, psychologists can propose new theories, conduct studies, and describe results (with potential benefits to countless people) by employing the scientific method. Sometimes theories that are proposed to explain human behavior are quite abstract and even strange. Observing the world around them, scientists use their critical thinking and skepticism to devise explanations for a wide range of human behavior. These theories are sometimes counterintuitive, meaning that they run counter to common expectations or previously held notions. An example may help to illustrate how even abstract theory can be explored through empirical research.

In the 1970s anthropologist Ernest Becker (1972) drew together theory and research from a broad array of social sciences to devise a grand theory of human life and culture. According to Becker, an important human characteristic that has evolved over the centuries is our amazing intellectual capacity. One result of this capacity is that unlike other animals, we humans are aware of our own vulnerability, notably the reality of our own deaths. This awareness of our mortality creates the potential for overwhelming terror. Yet somehow we manage to go about our daily lives without being preoccupied by the terrifying reality of death. Why is this so?

According to Becker, as our intellectual capacity evolved, so did our capacity to create and invest in culture. Culture provides the customary beliefs, practices, religious rules, and social order for humans living together. People in the same culture often share attitudes, values, and goals. Our culture gives us the framework to understand what behavior is appropriate and what is not. Culture provides answers to questions such as, How many wives should a man have? and Should children work to support their family?

Becker asserted that being part of a larger culture shields us from the terror of our own mortality. He maintained that by investing in our cultural worldview (our beliefs, routine practices, and standards for conduct), we are able to enjoy real and symbolic immortality. Real immortality is provided by religious ideas about life after death. Symbolic immortality derives from our ability to contribute to a culture that will outlive us. As long as we feel that we are valued members of a culture, this status will buffer us against our fears of personal death. Becker's theory is known as terror management theory (TMT) (Solomon, Greenberg, & Pyszczynski, 1991).

TMT theory is abstract and not very intuitive. It might strike you as “out there.” You might conclude that it is a theory that could not possibly be studied scientifically. How could one actually develop operational definitions for the variables in Becker's model? In fact, TMT has led to a number of provocative laboratory investigations that support Becker's views. Using the scientific method, social psychologists Jeff Greenberg, Sheldon Solomon, and Tom Pyszczynski (1997) derived some specific hypotheses from Becker's broad theory. One such hypothesis is that when people are reminded of their own death, we would expect them to show a tendency to champion their cultural worldview. That is, when our own death is made salient (real) to us, we should be more likely to defend ourselves against mortality by investing strongly in our cultural worldview. Thus, the awareness of death ought to lead to worldview defense.

Becker asserted that being part of a larger culture shields us from the terror of our own mortality.

How might we study this prediction empirically using an experimental design? First the variables must be operationalized. In this case, the independent variable (the cause) is death awareness. To make people more aware of their own deaths, these researchers asked participants to take a few minutes and to write a description of their own deaths—to describe what would happen to them physically and emotionally when they die (Arndt & others, 2005). This “mortality salience” condition is the experimental condition. The dependent variable in our prediction is the defense of one's worldview (the effect). How might this dependent variable be operationalized? One way that worldview defense might show itself is in attitudes about people who behave in ways that contradict the cultural worldview of what is appropriate—for example, criminals. In an early study, researchers asked a sample of 22 municipal court judges either to write about their own deaths or not to do this assignment, and then presented all of them with the same hypothetical case report of a woman arrested for prostitution (Rosenblatt & others, 1989). The judges were asked to set bail for the woman. Judges who had written about their own deaths gave the woman a much higher bond ($455 versus $50). In subsequent studies, these researchers and others have shown that reminding people of their own death tends to increase the tendency to judge harshly individuals who defy our cultural worldview and also to increase our own self-esteem.

Still, this research might seem artificial. How often do you sit down and write about your own death, and in this way get a reminder of your own mortality? Is this work ecologically valid—that is, does it represent how things work in the real world? Although you may not write such descriptions, reminders of death are in fact quite common: the violence we view on TV and in movies, news reports on war and acts of terrorism, the deaths of acquaintances, and even the cemetery or funeral home we drive by on our way home. These common experiences might serve as natural sources of mortality concern.

The terrorist attacks of September 11 might be thought of as a strong mortality salience manipulation. That is, thinking about 9/11 might cause individuals to feel a great deal of death anxiety. And from Becker's theory we would predict that thinking about 9/11 might require individuals to bolster their cultural worldview, just like the judges who were harsher toward a prostitute after thinking about their own deaths. Research has shown that reminders of September 11, such as the haunting images of the jet planes hitting the Trade Center towers, make death thoughts more accessible. For example, in one study, death accessibility (the dependent variable) was operationalized by having participants complete word fragments (Landau & others, 2004). After seeing reminders of 9/11, participants were more likely to complete the fragment COFF_ _ as coffin rather than coffee. Can such naturally occurring reminders of death influence our worldview in ways that matter?

Page 49

A series of studies by Mark Landau and colleagues (2004) revealed how reminders of death can cause changes in political allegiances. This research was conducted prior to the November 2004 presidential election when the two main candidates for president were President George W. Bush and Senator John Kerry. In this research, students who were randomly assigned to complete the mortality salience condition (that is, write about their own deaths) or to write about the attacks of 9/11 were more likely than students who had written, for the same amount of time, about the experience of dental pain (a control condition meant to hold anxiety constant) to express increasing support for George W. Bush. That is, students who had written about their own deaths or the events of 9/11 expressed more favorable attitudes toward President Bush compared to those who wrote about dental pain.

In a final study, participants who wrote about their own mortality showed increases in their favorability ratings of George W. Bush and less favorable ratings of John Kerry, and they judged themselves as more likely to vote for Bush than Kerry (Landau & others, 2004). These provocative findings suggest that unconscious concerns about death heighten the appeal of a charismatic leader. It is interesting to note that on October 29, just prior to the 2004 presidential election, Osama bin Laden (certainly a powerful reminder of the events of 9/11) appeared in a video criticizing George W. Bush. Bush later noted that he felt bin Laden's speech only helped him win the election (Reuters, 2006). Indeed, he may have been correct if bin Laden's appearance heightened death concerns for U.S. viewers.

Terror management theory and research provide a powerful example of how broad theory can be translated into strong empirical research. This theory has also yielded surprising findings with regard to more positive human characteristics. Concerns about death can lead either to our becoming narrow and defensive or, conversely, to our becoming more creative and more concerned about leaving a good legacy for the future (Routledge, Arndt, & Sheldon, 2004). Relationships, religious faith, creativity, and sharing our values with others all have been shown to diminish the need to engage in worldview defense when individuals are reminded of their mortality (Jonas & Fischer, 2006; Mikulincer, Florian, & Hirschberger, 2004; Routledge, Arndt, & Sheldon, 2004). In this way, concern over mortality can be a strong motivator to contribute positively to the world.

A final caution is worth noting in interpreting the results of an experiment. Even if the design was solid and free of obvious confounds or biases, some uncertainty may remain about precisely what aspect of the experimental manipulation caused the results in the dependent measure. An example is provided by the fascinating body of research on expressive writing begun by James Pennebaker. He and his colleagues (Pennebaker & Chung, 2007) conducted a number of studies that converge on the same conclusion: Writing about your deepest thoughts and feelings concerning your most traumatic life event leads to a number of health and well-being benefits.

In these studies, each participant is randomly assigned to write about one of two topics—either the individual's most traumatic life event or a relatively uninteresting topic (for example, his or her plans for the day). Assignment of the specific topic is meant to control for the act of writing itself (Pennebaker & Graybeal, 2001). The participants write about the same topic for 3 or 4 consecutive days for about 20 minutes each day. Weeks or months after writing, participants in the trauma writing group have better physical health than those in the control group. Since the first traumatic writing study, a host of researchers have replicated these effects, showing that writing about trauma is associated with superior immune function, better response to a vaccine, higher psychological well-being, better adjustment to coming to college, and more quickly finding employment after being laid off from work (Lepore & Smyth, 2002). Thus, we might conclude that documenting one's deepest thoughts and feelings about traumatic life events is necessary to attain what has been termed the “healing power” of writing.

Page 50

Note, however, that the participants in the trauma group were not only writing about a trauma. They were also documenting an important personal experience. Is it necessary to focus on a trauma to benefit from writing? Might there not be other, less negative aspects of life that are equally meaningful and that might bring health benefits when they are the subject of personal writing? Indeed, researchers recently have begun to examine the impact of writing about a variety of topics for health and well-being. For example, research has shown that writing only about the benefits of a traumatic life event—how a person has grown or become a better person because of the event—also leads to health benefits (King & Miner, 2000; Low, Stanton, & Danoff-Burg, 2006).

In addition, writing about one's life dreams not only produced health benefits equal to writing about a traumatic life event but did so while also boosting positive mood (King, 2001). In one study, writing about one's most intensely positive experience also led to health benefits (Burton & King, 2004). These findings and others have prompted researchers to rethink what the mechanisms underlying writing benefits might be (King, 2002). It may be that writing about meaningful and important life experiences is what is needed to benefit from writing, regardless of whether these experiences are negative or positive. We will return to the power of writing for health and wellness in Chapter 12.

At this point, you have read about several different types of research in psychology. For another look at how these research methods differ, see Figure 2.4. And to read about how psychologists have used the experimental research method to translate a broad theory into testable findings about terrorism, death anxiety, and political allegiance, see the Intersection.

Psychology's Research Methods Applied to Dreaming
Psychologists can apply very different methods to study the same phenomenon. Notice here how the object of study, dreams, can influence the usefulness of various methods.