Chapter 2: Studying Behavior Scientifically PDF

pas82760_ch02_027-059 8/10/07 12:19 PM Page 27 Techbooks (PPG-Quark) CHAPTER TWO Studying Behavior Scientifically 2 CHAPTER OUTLINE SCIENTIFIC PRINCIPLES IN PSYCHOLOGY THRE...

pas82760_ch02_027-059 8/10/07 12:19 PM Page 27 Techbooks (PPG-Quark) CHAPTER TWO Studying Behavior Scientifically 2 CHAPTER OUTLINE SCIENTIFIC PRINCIPLES IN PSYCHOLOGY THREATS TO THE VALIDITY OF RESEARCH Scientific Attitudes Confounding of Variables Gathering Evidence: Steps in the Scientific Process Placebo Effects Two Approaches to Understanding Behavior Experimenter Expectancy Effects Defining and Measuring Variables Replicating and Generalizing the Findings BENEATH THE SURFACE Science, Psychics, and the ETHICAL PRINCIPLES IN RESEARCH Paranormal Ethical Standards in Human Research Ethical Standards in Animal Research ANALYZING AND INTERPRETING DATA Being a Smart Consumer of Statistics METHODS OF RESEARCH Using Statistics to Describe Data Descriptive Research: Recording Events Using Statistics to Make Inferences WHAT DO YOU THINK? Should You Trust Internet and Pop Meta-Analysis: Combining the Results of Many Media Surveys? Studies Correlational Research: Measuring Associations between Events CRITICAL THINKING IN SCIENCE AND EVERYDAY LIFE RESEARCH CLOSE-UP Very Happy People APPLYING PSYCHOLOGICAL SCIENCE Evaluating Claims in WHAT DO YOU THINK? Does Eating Ice Cream Cause Research and Everyday Life People to Drown? Experiments: Examining Cause and Effect 27 pas82760_ch02_027-059 8/10/07 12:20 PM Page 28 Techbooks (PPG-Quark) I have no special talents. I am only passionately curious. —ALBERT EINSTEIN inter was around the corner, and Teryl’s Sunday-morning drive to Oregon City was about to take W an unexpected twist. The pickup truck in front of her hit a slick spot, veered off the road, and went over the edge. Teryl, who has paralyzed legs and only partial use of her arms, stopped her minivan, low- ered herself into her wheelchair, and headed to the crash site. She then got out of her wheelchair, slid down the wet embankment, crawled to reach the bleeding, dazed driver, and administered first aid un- til the paramedics arrived. Said Teryl, “I think anybody would have done that. You see a car go down a ditch, and I can’t imagine not stopping to help” (Seattle Times, 1997, December 11). Yet unfortunately, as we shall see over and over again, the unimaginable does indeed occur.  n a warm summer evening in July 2006, the sound of gunshots inside a home shatter the quiet of O a normally peaceful North Toronto neighborhood. About 20 people hear the shots, but “residents go about their business walking the dog, watering flowers, relaxing in the backyard” (Huffman, 2006, August 2). No one even calls the police. Eventually, two people are found dead inside the garage, dis- covered by the wife of one of the victims. Ten months later, in Detroit, a 91-year-old man is hospitalized after being severely beaten by a carjacker. Bystanders witness the beating, but fail to intervene. A store surveillance camera captures the incident on video. The victim’s son remarks in disbelief after viewing the video, “ I’ll never get over those other guys standing around. I never thought I’d see anything like that” (Schmitt, 2007, May 15). These and other similar incidents spark the memory of the infamous case of Kitty Genovese, who in March 1964 was attacked by a knife-wielding assailant as she returned from work to her New York City apartment. The attack occurred around 3 A.M. and lasted about 30 minutes, during which time dozens of neighbors saw her being stabbed or heard Genovese’s screams and pleas for help. Many went to their windows to find out what was happening. Yet nobody assisted her, and by the time the police arrived, she had died. The incident drew international attention from a shocked public, and commentators expressed outrage over “bystander apathy” and people’s refusal to “get involved.” Science frequently has all the mystery of a good science. These principles also promote a way of detective story. Consider the psychological puz- thinking—critical thinking—that can serve you zle of bystander intervention. If you were in well in many aspects of your life. trouble and needed help from bystanders, would you receive it? Ordinary citizens like Teryl often act decisively to help someone in SCIENTIFIC PRINCIPLES need (Figure 2.1). But, as the Kitty Genovese mur- IN PSYCHOLOGY der and similar tragedies illustrate, people do not always come to the aid of others. Why do by- At its core, science is an approach to asking and standers sometimes risk injury and death to assist answering questions about the universe around a stranger yet at other times fail to intervene— us. Certainly, there are other ways we learn about even when helping or calling the police entails our world and ourselves: through reason, intu- little personal risk? We will return to this puzzle ition, and common sense; religion and spirituality; FIGURE 2.1 shortly. the arts; and the teachings of family, friends, and What determines whether a In this chapter we explore principles and others. What distinguishes science from these ap- bystander will help a victim? methods that form the foundation of psychological proaches is a process guided by certain principles. 28 pas82760_ch02_027-059 8/10/07 12:20 PM Page 29 Techbooks (PPG-Quark) STUDYING BEHAVIOR SCIENTIFICALLY 29 SCIENTIFIC ATTITUDES form a hypothesis. Noting that each bystander  Focus 1 probably knew that other bystanders were also Describe three key scientific Curiosity, skepticism, and open-mindedness are witnessing Kitty Genovese’s plight, Darley and attitudes and how they guided driving forces behind scientific inquiry. Like a Darley and Latané’s response to Latané proposed that a diffusion of responsibility the Genovese murder. child who constantly asks “Why?”, the good sci- reduced the likelihood that any one bystander entist is intensely curious. And like a master de- would intervene. This tentative explanation is tective, the good scientist is an incurable skeptic. then translated into a hypothesis, a specific predic- Each claim is met with the reply, “Show me your tion about some phenomenon that often takes the evidence.” Scientists also must remain open- form of an “If-Then” statement: “In an emergency, minded to conclusions supported by facts, even if IF multiple bystanders are present, THEN the those conclusions refute their own beliefs. likelihood that any one bystander will intervene is Following the Kitty Genovese murder, two reduced.” psychology professors in New York City, John Darley and Bibb Latané, met for dinner. They were Step 3: Test Hypothesis by Conducting Research. so curious about how dozens of people could wit- The third step is to test the hypothesis by con- ness a violent crime and not even call the police ducting research. Darley and Latané (1968) staged that they decided to investigate further. Darley an “emergency” in their laboratory and recorded and Latané also were skeptical of the “bystander people’s responses. Male undergraduate partici- apathy” explanation offered by the media; they be- pants were told that they would be discussing lieved it unlikely that all of the bystanders could “college experiences.” To ensure privacy, they have been apathetic. They noted that the by- would be in separate rooms, would communicate standers could see that other neighbors had turned through an intercom system, and the experi- on their lights and were looking out their win- menter would not listen to their conversation. The dows. Each bystander might have been concerned students understood that they would take turns about Kitty Genovese’s plight but assumed that speaking for several rounds. In each round, a stu- someone else surely would help or call the police. dent would have 2 minutes to speak, during Darley and Latané reasoned that the presence which time the others would be unable to inter- of multiple bystanders produced a diffusion of rupt or be heard, because their microphones responsibility, a psychological state in which each would be turned off. person feels decreased personal responsibility for As the discussion began over the intercom, a intervening. They performed several experiments speaker described his difficulties adjusting to col- to test their explanation but had to remain open- lege life and disclosed that he suffered from minded to the possibility that the findings would seizures. During the next round, this same not support their point of view. speaker began to gasp, saying: “‘... Could some- body-er-er—help... [choking sounds]... I’m gonna die-er-er—I’m gonna die-er—help... GATHERING EVIDENCE: STEPS IN THE seizure’ [chokes, then silence]” (Darley & Latané, SCIENTIFIC PROCESS 1968, p. 379). Science involves a continuous interplay between Unbeknownst to the students, they were ac-  Focus 2 observing and explaining events. Figure 2.2 shows tually listening to a tape recording. This ensured Use Darley and Latané’s research the following five steps that reflect how scientific that all of them were exposed to the identical or another study to illustrate five “emergency.” To test how the number of by- major steps in the scientific inquiry often proceeds. process. standers influences helping, Darley and Latané Step 1: Identify a Question of Interest. Curiosity assigned students to one of three conditions on a sparks the first step of scientific inquiry: identify- random basis. Each student actually was alone ing a question of interest. From personal experi- but was led to believe that, on the intercom sys- ences, news events, scientific articles and books, tem, (1) he was alone with the victim, (2) there and other sources, scientists observe something was another listener present, or (3) there were that piques their interest, and they ask a question four other listeners present. The students be- about it. Darley and Latané observed that nobody lieved that the seizure was real and serious. But helped Kitty Genovese and then asked the ques- did they help? tion “Why?”. Step 4: Analyze Data, Draw Tentative Conclusions, Step 2: Gather Information and Form Hypothesis. and Report Findings. At the fourth step, re- Next, scientists examine whether any studies, the- searchers analyze the information (called data) ories, and other information already exist that they collect, draw tentative conclusions, and re- might help answer their question, and then they port their findings to the scientific community. As pas82760_ch02_027-059 8/10/07 3:46 PM Page 30 107:MHIY036:pas82760_ch02: 30 CHAPTER 2 USING THE SCIENTIFIC METHOD Examining bystander intervention: Why do people sometimes fail to help a victim in need during an emergency, even when there is little or no personal risk? What factors increase or decrease the likelihood that a bystander will 1 intervene? STEP IDENTIFY 2 Identify Question of Interest Kitty Genovese is murdered. ? The attack lasts over 30 minutes. Neighbors fail even to call the STEP HYPOTHESIZE police until it is too late. The public is shocked. Why did no one help? Gather Information and Form Hypothesis 3 A diffusion of responsibility may have occurred. Hypothesis: IF multiple bystanders are present, THEN each bystander’s likelihood of intervening will decrease. STEP TEST Test Hypothesis by Conducting Research Conduct an experiment by creating an emergency in a controlled set- ting. Manipulate (control) the number 4 A of other bystanders that each partici- B pant believes to be present, and then measure whether and how quickly each participant helps the victim. STEP ANALYZE 5 Analyze Data, Draw Tentative Conclusions, and Report Findings The data reveal that helping decreases as the BUILD perceived number of bystanders increases. STEP The hypothesis is supported. (If the data are Build a Body of Knowledge; Ask Further found not to support the hypothesis, revise Questions; Conduct More Research; hypothesis or procedures and retest.) Develop and Test Theories Additional experiments support the hypoth- esis. A theory of social impact is developed based on these findings. The theory is then tested directly by deriving new hypotheses and conducting new research. FIGURE 2.2 Using the scientific method. pas82760_ch02_027-059 8/10/07 12:20 PM Page 31 Techbooks (PPG-Quark) STUDYING BEHAVIOR SCIENTIFICALLY 31 (Latané & Bourgeois, 2001). Scientists use the the- Cumulative proportion helping (%) Participant is the 100 only bystander ory to formulate new hypotheses, which are then tested by conducting still more research. In this 80 Participant plus manner, the scientific process becomes self- 1 other bystander correcting. If research consistently supports the hy- 60 potheses derived from the theory, confidence in the theory becomes stronger. If the predictions made 40 Participant plus by the theory are not supported, then it will need to 4 other bystanders be modified or, ultimately, discarded. 20 0 60 120 180 240 TWO APPROACHES TO UNDERSTANDING Time from beginning of seizure BEHAVIOR (seconds) Humans have a strong desire to understand why FIGURE 2.3 things happen. Why do scientists favor the pre- Helping in an emergency. Participants who believed that they ceding step-by-step approach to understanding were the only bystander who could hear a seizure victim’s plea for behavior over the approach typically involved in help were more likely to take action than were participants who everyday commonsense: hindsight? believed that either one or four additional bystanders were listening. SOURCE: Data from Darley & Latané, 1968. Hindsight (After-the-Fact Understanding) Figure 2.3 shows, Darley and Latané found that Many people erroneously believe that psychology  Focus 3 all participants who thought they were alone with is nothing more than common sense. “I knew that Explain the major drawback of hindsight understanding. What the victim helped within 3 minutes of the seizure. all along!” or “They had to do a study to find that approach to understanding do As the number of presumed bystanders increased, out?” are common responses to some psychologi- scientists prefer? Why? the proportion of actual participants who helped cal research. For example, decades ago a New York decreased, and those who helped took longer to Times book reviewer criticized a report titled The respond. These findings support the diffusion-of- American Soldier (Stouffer et al., 1949a, 1949b), responsibility explanation and illustrate how re- which summarized the results of a study of the search can contradict commonsense adages such attitudes and behavior of U.S. soldiers during as “There’s safety in numbers.” World War II. The reviewer blasted the govern- Darley and Latané then submitted a report de- ment for spending a lot of money to “tell us noth- scribing their research to a scientific journal. ing we don’t already know.” Expert reviewers favorably judged the quality and Consider the following statements. How importance of the research, so the journal pub- would you account for each of them? lished the article. Publishing research is essential 1. Compared to White soldiers, Black soldiers to scientific progress. It allows fellow scientists to were less motivated to become officers. learn about new ideas and findings, to evaluate the research, and to challenge or expand on it. 2. During basic training, soldiers from rural areas had higher morale and adapted better Step 5: Build a Body of Knowledge. At the fifth than soldiers from large cities. step, scientists build a body of knowledge about 3. Soldiers in Europe were more motivated to the topic in question. They ask further questions return home while the fighting was going (e.g., “What other factors affect bystander interven- on than they were after the war ended. tion?”), formulate new hypotheses, and test those hypotheses by conducting more research. As addi- You should have no difficulty explaining tional evidence comes in, scientists may attempt to these results. Typical reasoning might go some- build theories. A theory is a set of formal statements thing like this: (1) Due to widespread prejudice, that explains how and why certain events are related to Black soldiers knew that they had little chance of one another. Theories are broader than hypotheses. becoming officers. Why should they torment For example, dozens of experiments revealed that themselves wanting something that was unattain- diffusion of responsibility occurred across a range able? (2) It’s obvious that the rigors of basic train- of situations. Latané then combined the principle ing would seem easier to people from farm set- of diffusion of responsibility with other principles of tings, who were used to hard work and rising at group behavior to develop a broader theory of social the crack of dawn. (3) Any sane person would impact, which he and others have since used to have wanted to go home while bullets were flying explain a variety of human social behaviors and people were dying. pas82760_ch02_027-059 08/24/07 1:53 PM Page 32 Team B 107:MHIY036:pas82760_ch02: 32 CHAPTER 2 Did your explanations resemble these? If so, they are perfectly reasonable. There is one catch, however. The results of the actual study were the opposite of the preceding statements. In fact, Black soldiers were more motivated than White soldiers to become officers, city boys had higher morale than farm boys during basic training, and soldiers were more eager to return home after the war ended than during the fighting. When told these ac- tual results, our students quickly find explanations for them. In short, it is easy to arrive at reasonable after-the-fact explanations for almost any result. In everyday life, hindsight (after-the-fact ex- planation) is probably our most common method of trying to understand behavior. The Danish philosopher Søren Kierkegaard noted, “Life is lived forwards, but understood backwards.” The major limitation of relying solely on hindsight is that past events usually can be explained in many FIGURE 2.4 ways, and there is no sure way to know which—if The importance of testability. Is the scientist’s claim of discover- any—of the explanations is correct. Despite this ing an “eternal life potion” a testable hypothesis? Yes, because it is drawback, after-the-fact understanding can pro- possible to show the hypothesis to be false. If people drink it but vide insights and is often the foundation on which still die at some point in time, then we have refuted the hypothesis. further scientific inquiry is built. For example, Therefore it is testable. It is, however, impossible to absolutely prove Darley and Latané’s diffusion-of-responsibility ex- true. If a person drinks the potion, then no matter how long she or he lives—even a million years—she or he might die the next day. planation was initially based on after-the-fact rea- Thus, we cannot prove that the potion can make you live forever. soning about the Kitty Genovese murder. Copyright © 2004 by Sidney Harris: ScienceCartoonsPlus.com.  Focus 4 Reprinted with permission. Describe some characteristics of a good theory. Understanding through Prediction, Control, and Theory Building It is testable. It generates new hypotheses whose accuracy can be evaluated by gather- Whenever possible, scientists prefer to test their ing new evidence (Figure 2.4). understanding of “what causes what” more di- rectly. If we truly understand the causes of a given The predictions made by the theory are sup- behavior, then we should be able to predict the ported by the findings of new research. conditions under which that behavior will occur It conforms to the law of parsimony: If two in the future. Furthermore, if we can control those theories can explain and predict the same conditions (e.g., in the laboratory), then we should phenomenon equally well, the simpler the- be able to produce that behavior. ory is the preferred one. Darley and Latané’s research illustrates this Even when a theory is supported by many suc- approach. They predicted that due to a diffusion of cessful predictions, it is never regarded as an responsibility, the presence of multiple bystanders absolute truth. It is always possible that some fu- during an emergency would reduce individual ture observation will contradict it or that a newer helping. Next, they carefully staged an emergency and more accurate theory will take its place. The and controlled participants’ beliefs about the displacement of old beliefs and theories by newer number of bystanders present. Their prediction ones is the essence of scientific progress. was supported. Understanding through predic- Finally, although scientists use prediction as a tion and control is a scientific alternative to after- test of “understanding,” this does not mean that the-fact understanding. prediction requires understanding. Based on Theory building is the strongest test of scien- experience, even a child can predict that thunder tific understanding, because good theories gener- will follow lightning without knowing why it ate an integrated network of predictions. A good does so. But prediction based on understanding theory has several important characteristics. (i.e., theory building) has advantages: It satisfies It incorporates existing knowledge within a our curiosity and generates principles that can be broad framework; that is, it organizes infor- applied to new situations that we have not yet mation in a meaningful way. directly experienced. pas82760_ch02_027-059 8/10/07 12:20 PM Page 33 Techbooks (PPG-Quark) STUDYING BEHAVIOR SCIENTIFICALLY 33 LEVELS OF ANALYSIS Measuring Exam Stress Biological Psychological Environmental Stress-hormone levels measured at rest and General achievement anxiety measured by self- Aspects of immediate environment that create during an exam report personality test stress (e.g., difficulty of exam, time pressure, Measures of heart rate and respiration rate Preexam questionnaire ratings of worry, noise and heat levels) Physiological measures of muscle tension and tension, and anxiety Easy or difficult course grading standards set by sweating Behavioral observations of “nervous habits” instructor during exam (e.g., fingernail biting, foot Achievement expectations set by parents or wiggling, hair pulling) instructor Exam Stress FIGURE 2.5 Levels of analysis: Measuring exam stress. DEFINING AND MEASURING VARIABLES performance” could mean a single test score, a course grade, or one’s overall grade point aver- Psychologists study variables and the relations age. So, for our study, let’s define it as students’ among them. A variable, quite simply, is any final exam scores in an introductory chemistry characteristic or factor that can vary. Birth order is course. As for “stress,” before or during the exam a variable: Some people are first born, second we could measure students’ levels of muscle ten- born, and so on. People’s hair color, income, age, sion or stress hormones, or ask them to report sex, grade point average, and typing speed are how worried they feel. During the test we might variables: They vary from one person to another, observe their frequency of nail biting. We also and over time some also vary within a given could define stress in terms of environmental person. conditions, such as whether the exam questions Many variables that psychologists study rep- and grading scale are easy or difficult. Figure 2.5 resent abstract concepts that cannot be observed summarizes how we might operationally define directly. For example, “self-esteem,” “stress,” and exam stress at the biological, psychological, and “intelligence” are concepts that refer to people’s in- environmental levels. ternal qualities. We might say that Tyra has high Measurement is challenging because psy- self-esteem, Shaun is intelligent, and Claire feels chologists study incredibly varied and complex stressed, but how do we know this? We can’t di- processes. Some processes are directly observable, rectly look inside their heads and see “self-esteem,” but others are not. Fortunately, psychologists “intelligence,” or “stress,” yet such concepts must have numerous measurement techniques at their be capable of being measured if we are to study disposal (Figure 2.6). them scientifically. Because any variable may mean different Self-Reports and Reports by Others things to different people, scientists must define their terms clearly. And when conducting re- Self-report measures ask people to report on their  Focus 5 search, scientists must also define variables oper- own knowledge, attitudes, feelings, experiences, Why are operational definitions important? Identify five major ationally. An operational definition defines a vari- or behavior. This information can be gathered in ways to measure behavior and able in terms of the specific procedures used to produce several ways, such as through interviews or ques- explain a limitation of each one. or measure it. Operational definitions translate tionnaires. The accuracy of self-reports hinges on abstract concepts into something observable and people’s ability and willingness to respond hon- measurable. estly. Especially when questions focus on sensitive To illustrate, suppose we want to study the topics, such as sexual habits or drug use, self-re- relation between stress and academic perform- ports may be distorted by social desirability bias, ance among college students. How shall we the tendency to respond in a socially acceptable manner operationally define our variables? “Academic rather than according to how one truly feels or behaves. pas82760_ch02_027-059 8/10/07 12:20 PM Page 34 Techbooks (PPG-Quark) 34 CHAPTER 2 (a) (b) (c) FIGURE 2.6 (a) Self-report, (b) physiological, and (c) behavioral measures are important scientific tools for psychologists. Researchers try to minimize this bias by establish- parent and child jointly perform a task, we ing rapport with participants and allowing them might code the parent’s behavior into categories to respond confidentially or anonymously. Ques- such as “praises child,” “assists child,” and tionnaires can also be designed to reduce social de- “criticizes child.” Observers must be trained to sirability bias. use the coding system properly so that their We also can get information about someone’s measurements will be reliable (i.e., consistent). If behavior by obtaining reports made by other people, two observers watching the same behaviors re- such as parents, spouses, and teachers who know peatedly disagree in their coding (e.g., one says the person. College students might be asked to the parent “praised” and another says the par- rate their roommates’ personality traits, and job ent “assisted”), then the data are unreliable and supervisors might be asked to rate a worker’s of little use. competence. As with self-reports, researchers try Humans and other animals may behave dif- to maximize participants’ honesty in reporting ferently when they know they are being ob- about other people. served. To counter this problem, researchers may camouflage themselves or use unobtrusive measures, which record behavior in a way that keeps Measures of Overt Behavior participants unaware that they are being observed. Another measurement approach is to record overt To illustrate, scientists from the Centers for (i.e., directly observable) behavior. In an experiment Disease Control assessed the effects of a “safer on learning, we might measure how many sex” program by counting the number of used errors a person or laboratory rat makes while condoms that turned up in a Baltimore sewage performing a task. In an experiment on drug treatment plant before and after the program effects, we might measure people’s reaction was begun. (We never said that science is always time—how rapidly they respond to changing glamorous.) stimulus conditions (such as the onset or offset Psychologists also gather information about of a light)—after ingesting various amounts of behavior by using archival measures, which are alcohol. In Darley and Latané’s (1968) bystander records or documents that already exist. For example, emergency experiment, they recorded whether researchers assessing a program to reduce drunk and how quickly college students helped a seizure driving could examine police records to measure victim. how many people were arrested for driving while Psychologists also develop coding systems to intoxicated before and after the program was im- record different categories of behavior. While a plemented. pas82760_ch02_027-059 8/10/07 12:20 PM Page 35 Techbooks (PPG-Quark) STUDYING BEHAVIOR SCIENTIFICALLY 35 Psychological Tests planation and a testable hypothesis, (3) conduct- Psychologists develop and use specialized tests to ing research to test the hypothesis, (4) analyzing measure many types of variables. For example, the data, drawing a tentative conclusion, and re- personality tests, which assess people’s personality porting one’s findings to the scientific community, traits, often contain a series of questions that ask and (5) building a body of knowledge by asking how a person typically feels or behaves (e.g., further questions, conducting more research, and “True or False: I prefer to be alone rather than in developing and testing theories. social gatherings.”). In essence, such tests are spe- cialized self-reports. Other personality tests pres-  In everyday life, we typically use hindsight to ex- ent a series of ambiguous stimuli (e.g., pictures plain behavior. Hindsight is flawed because there that could have different meanings), and person- may be many possible explanations and no way to ality traits are judged based on how a person assess which is correct. Psychologists prefer to interprets these stimuli. test their understanding through prediction, con- Other psychological tests consist of perform- trol, and theory building. ance tasks. For example, intelligence tests may ask  A good theory organizes known facts, gives rise to people to assemble objects or solve arithmetic additional hypotheses that are testable, is sup- problems. Neuropsychological tests help diagnose ported by the findings of new research, and is normal and abnormal brain functioning by meas- parsimonious. uring how well people perform mental and phys-  An operational definition defines a concept or ical tasks, such as recalling lists of words or variable in terms of the specific procedures used manipulating objects (Holtzer et al., 2005). to produce or measure it.  To measure behavior, psychologists obtain Physiological Measures people’s self-reports and reports from others who Psychologists also record physiological responses know the participants, directly observe behavior, to assess what people are experiencing. Measures use unobtrusive measures, analyze archival data, of heart rate, blood pressure, respiration rate, hor- administer psychological tests, and measure monal secretions, and electrical and biochemical physiological responses. processes in the brain have long been the mainstay of researchers working within the biological per- spective, but these measures have become increas- ingly important in many other areas of psychology. ETHICAL PRINCIPLES Physiological responses can have their own IN RESEARCH interpretive problems, the main one being that we don’t always understand what they mean. For ex- When designing their research, psychologists  Focus 6 ample, if a person shows increased heart rate and must weigh the knowledge and possible applica- Identify major ethical principles and standards in human and brain activity in a particular situation, what emo- tions to be gained against potential risks to re- animal research. tion or thought is being expressed? The links search participants. To safeguard the rights of par- between specific patterns of physiological activity ticipants, researchers must adhere to ethical and particular mental events are far from being standards set by government regulations and na- completely understood. tional psychological organizations. Animal sub- In sum, psychologists can measure behavior jects must also be treated in accord with estab- in many ways, each with advantages and disad- lished ethical guidelines. At academic and vantages. To gain greater confidence in their find- research institutions, special committees review ings, researchers may use several types of meas- the ethical issues involved in research proposals. ures within a single study. If a proposed study is considered ethically ques- tionable, it must be modified or the research can- not be conducted. The Ethics Code of the American Psycholog- IN REVIEW ical Association (APA) was first published in 1953. The current code (APA, 2002) builds on  Curiosity, skepticism, and open-mindedness are work by international and national commis- key scientific attitudes. The scientific process sions charged with developing ethical guide- proceeds through several steps: (1) identifying a lines for biomedical and behavioral research. question of interest, (2) formulating a tentative ex- The APA code sets forth five broad ethical prin- ciples that represent ideals toward which all pas82760_ch02_027-059 8/10/07 12:20 PM Page 36 Techbooks (PPG-Quark) 36 CHAPTER 2 FIGURE 2.7 Ethical standards are designed to protect the welfare of humans and nonhumans in psychological research. psychologists should strive. These principles potential risks must be identified and then include: weighed against the potential benefits to be gained. When children, seriously disturbed Beneficence: seeking to benefit other people mental patients, or other people who cannot Responsibility: performing professional duties give true informed consent are involved, con- with utmost care sent must be obtained from their parents or Integrity: being honest and accurate guardians. To safeguard a participant’s right to Justice: enhancing all people’s access to the privacy, researchers typically gather and report benefits of psychological knowledge data in ways that keep participants’ identity Respect: respecting people’s dignity and rights anonymous or confidential. to confidentiality and self-determination Deception, which occurs when participants (APA, 2002) are misled about the nature of a study, is contro- versial. Consider the Darley and Latané (1968) bystander experiment. Participants were not ETHICAL STANDARDS IN told that the study was going to examine how HUMAN RESEARCH they would respond to an emergency, nor were they informed that the procedure (someone The APA’s Ethics Code also provides dozens of presumably having a seizure) might cause them specific guidelines for psychological activities, stress. including research (Figure 2.7). According to the Deception violates the principle of informed ethical standard of informed consent, before peo- consent, but its proponents argue that when ple agree to participate in research, they should be studying certain types of behaviors, deception is informed about: the only way to obtain natural, spontaneous the study’s purpose and procedures; responses from participants. Darley and Latané’s participants, for example, had to believe that the the study’s potential benefits; emergency was significant and real. potential risks to participants; Guidelines currently permit deception only the right to decline participation and withdraw at when no other feasible alternative is available any time without penalty; and the study has scientific, educational, or ap- whether responses will be confidential and, if not, plied benefits that clearly outweigh the ethical how privacy will be safeguarded. costs of deceiving participants. When deception is used, the true purpose of the study should be The principle of informed consent strongly explained to participants after it is over. The emphasizes the importance of risk/benefit overwhelming majority of psychological studies analysis. In other words, a proposed study’s do not involve deception. pas82760_ch02_027-059 8/10/07 12:20 PM Page 37 Techbooks (PPG-Quark) STUDYING BEHAVIOR SCIENTIFICALLY 37 ETHICAL STANDARDS IN animals from being shot and killed (Gustavson ANIMAL RESEARCH & Gustavson, 1985). Although animal research has declined According to the APA’s Committee on Animal slightly in recent decades, the ethical questions re- Research and Ethics (CARE, 2005), animals are main as vexing as ever. What is most encouraging subjects in perhaps 7 to 8 percent of psychologi- is that the welfare of animals in research is receiv- cal studies. This includes research done in the ing the careful attention it deserves. wild and in more controlled settings. Rodents and birds comprise 90 percent of the animals studied; nonhuman primates comprise another 5 percent. IN REVIEW Some psychologists study animals to dis- cover principles that shed light on human behav-  Psychological research follows extensive ethical ior, and some do so to learn more about other guidelines. In human research, key issues are the species. National surveys find that the vast major- use of informed consent, the participants’ right to ity of psychologists and college psychology ma- privacy, potential risks to participants, and the use jors believe that animal research is necessary for of deception. scientific progress in psychology (Plous, 1996a,  Ethical guidelines require that animals be treated 1996b). As in medical research, however, some humanely and that the risks to which they are studies expose animals to conditions considered exposed be justified by the potential importance too hazardous for humans. of the research. As in human research, before APA and federal government guidelines re- animal research can be conducted it must be quire that animals be treated humanely and that reviewed and approved, often by ethics review the potential importance of the research clearly boards that include nonscientists. justifies the risks to which they are exposed. This determination, however, is not always easy to make, and people of good will can disagree. For example, should researchers be allowed to inject a METHODS OF RESEARCH drug into an animal in order to learn whether that drug might permanently impair memory? Before Like detectives searching for clues to solve a case, animal research can be conducted, it must be re- psychologists conduct research to gather evidence viewed and approved by panels that often in- about behavior and its causes. The research method clude nonscientists. chosen depends on the problem being studied, the Animal research is debated both outside and investigator’s objectives, and ethical principles. within the psychological community (Herzog, 2005). Psychologists agree that it is morally wrong to subject animals to needless suffering. DESCRIPTIVE RESEARCH: RECORDING Many scientists, however, do not agree with the EVENTS former head of the American Anti-Vivisection The most basic goal of science is to describe phe-  Focus 7 Society who maintained that animals should nomena. In psychology, descriptive research seeks Discuss three types of descrip- never be used in research “which is not for the to identify how humans and other animals behave, par- tive research, and explain the ad- benefit of the animals involved” (Goodman, vantages and disadvantages of ticularly in natural settings. Such research provides each. 1982, p. 61). Proponents point to numerous im- valuable information about the diversity of be- portant medical and psychological advances havior and may yield clues about potential cause- made possible by animal research. They ask, effect relations that are later tested experimen- “Does the prospect of finding a cure for cancer tally. Case studies, naturalistic observation, and or of identifying harmful drug effects or the surveys are research methods commonly used to causes of psychological disorders justify expos- describe behavior. ing some animals to harm?”. Proponents also point to examples of how Case Studies: The Hmong Sudden animal research has benefited animals them- Death Syndrome selves. For example, using learning principles discovered in studies with dogs, researchers A case study is an in-depth analysis of an individ- have changed the behavior of coyotes, bears, ual, group, or event. By studying a single case in and other wild animals that were endangering great detail, the researcher typically hopes to humans or livestock, thereby sparing those wild discover principles of behavior that hold true for pas82760_ch02_027-059 8/10/07 12:20 PM Page 38 Techbooks (PPG-Quark) 38 CHAPTER 2 people or situations in general. Data may be first night he woke suddenly, short of breath, gathered through observation, interviews, psy- from a dream in which a cat was sitting on his chological tests, physiological recordings, or chest. The second night... a figure, like a large task performance. black dog, came to his bed and sat on his chest... One advantage of a case study is that when a and he grew quickly and dangerously short of rare phenomenon occurs, this method enables sci- breath. The third night, a tall, white-skinned fe- entists to study it closely. A second advantage is male spirit came into his bedroom... and lay on that a case study may challenge the validity of a top of him. Her weight made it increasingly diffi- theory or widely held scientific belief. Perhaps the cult for him to breathe.... After 15 minutes, the biggest advantage of a case study is that it can be a spirit left him and he awoke, screaming. (Tobin & vibrant source of new ideas that may subsequently Friedman, 1983, p. 440) be examined using other research methods. Con- sider the following example of a case study. Vang’s report attracted scientific interest be- Vang is a former Hmong (Laotian) soldier cause about 25 Laotian refugees in the United who resettled in Chicago in 1980 after escaping States already had died of what was termed the the ravages of war in Laos. Vang had traumatic “Hmong sudden death syndrome.” The cases memories of wartime destruction and severe guilt were similar to Vang’s: A healthy person died in about leaving his brothers and sisters behind his or her sleep after exhibiting labored breathing, when he fled with his wife and child (Figure 2.8). screaming, and frantic movements. The U.S. Cen- The culture shock created by moving from rural ters for Disease Control investigated these deaths Laos to urban Chicago increased Vang’s stress. and concluded that the deaths were triggered by a According to a mental health team, Vang experi- combination of the stress of resettlement, guilt enced problems almost immediately: over abandoning family in Laos, and the Hmong’s [He] could not sleep the first night in the apart- cultural beliefs about angry spirits. ment, nor the second, nor the third. After three The authors of Vang’s case study concluded days... Vang came to see his resettlement that he might have been a survivor of the sud- worker... Moua Lee. Vang told Moua that the den death syndrome. The role of cultural beliefs is suggested by what happened next. Vang went for treatment to a Hmong woman regarded as a shaman (a person, acting as both doctor and priest, who is believed to work with spirits and the supernatural). She told him his problems were caused by unhappy spirits and performed ceremonies to release the spirits. Vang’s night- mares and breathing problems during sleep ceased. Vang’s case study suggests that cultural be- liefs and stress may profoundly influence physi- cal well-being. This work was followed by other studies of Hmong immigrants and stimulated in- terest in the relation between cultural beliefs and health (Miller & Rasco, 2004). The major limitation of a case study is that it is a poor method for determining cause-effect relations. In most case studies, explanations of behavior occur after the fact and there is little opportunity to rule out alternative explanations. The fact that Vang’s symptoms ended after seeing a shaman might not have had anything to do with his cultural beliefs; it could have been pure coinci- dence, or other changes in Vang’s life could have FIGURE 2.8 been responsible. Many Hmong refugees who escaped the ravages of war in their A second potential drawback concerns the homeland experienced great stress and guilt when they resettled in North America. This stress, combined with cultural beliefs about generalizability of the findings: Will the principles angry spirits, may have contributed to the Hmong sudden death uncovered in a case study hold true for other syndrome, which eventually claimed more than 40 lives. people or in other situations? The question of pas82760_ch02_027-059 8/10/07 12:20 PM Page 39 Techbooks (PPG-Quark) STUDYING BEHAVIOR SCIENTIFICALLY 39 generalizability pertains to all research methods, but drawing broad conclusions from a case study can be particularly risky. The key issue is the de- gree to which the case under study is representa- tive of other people or situations. A third drawback is the possible lack of ob- jectivity in the way data are gathered and inter- preted. Such bias can occur in any type of re- search, but case studies can be particularly worrisome because they are often based largely on the researcher’s subjective impressions. In sci- ence, a skeptical attitude requires that claims based on case studies be followed up by more comprehensive research methods before they are accepted. In everyday life, we should adopt a similarly skeptical view. When you encounter claims based on case examples or anecdotes, keep in mind that the case may be atypical or that the person making the claim may be biased. Try to seek out other evidence to support or re- fute the claim. Naturalistic Observation: Bullies in the Schoolyard In naturalistic observation, the researcher observes behavior as it occurs in a natural setting, and attempts to avoid influencing that behavior (Figure 2.9). This FIGURE 2.9 method is used extensively to study nonhuman Psychologists conduct naturalistic observations in many settings, animal behavior. For example, by observing including the classroom. African chimpanzees in the wild, British re- searcher Jane Goodall and other scientists found here are 3 of 10 categories representing different that chimpanzees display behaviors, such as intervention strategies: making and using tools, that were formerly be- lieved to lie only within the human domain Verbal Assertion: Verbally requesting that the (Goodall, 1986; Lonsdorf, 2006). bullying stop, without verbally attacking the Naturalistic observation is also used to study bully or victim (e.g., “Stop it,” “Knock it off.”). human behavior. Consider bullying at school, a Physical Assertion: Physically stepping in to topic that has received increasing attention from separate the bully and victim, but not physi- educators and psychologists (Kanetsuna et al., cally attacking either one. 2006). When you were a child, were you ever the Physical Aggression: Hitting, pushing, shov- victim of bullying? If so, did any schoolmates step ing, or otherwise physically engaging the in to help? In a three-year study, psychologists bully or victim. videotaped and audiorecorded the playground interactions of a sample of children during recess Overall, of the 306 bullying episodes ob- and lunch periods at two elementary schools in served, schoolmates were present 88 percent of Toronto (Hawkins et al., 2001). The researchers’ the time but intervened in only 19 percent of the main goal was to describe the nature of peer inter- episodes. In order, the three most common types ventions during episodes of schoolyard bullying. of intervention were verbal assertion alone, phys- When bullying occurs, how often do schoolmates ical aggression alone, and verbal assertion com- intervene? What strategies do they use? Are peer bined with physical assertion. interventions effective? Like case studies, naturalistic observation As in many observational studies, to answer does not permit clear conclusions about the these questions the researchers developed coding causal relations between variables. In the real systems so that the children’s behavior could be world, many variables simultaneously influence classified into meaningful categories. To illustrate, behavior, and they cannot be disentangled with pas82760_ch02_027-059 8/10/07 12:20 PM Page 40 Techbooks (PPG-Quark) 40 CHAPTER 2 this research technique. There also is the possibil- responses of an entire population of 25 million ity of bias in the way that researchers interpret American adolescents based on these data (U. S. the behaviors they observe. Finally, observers Census Bureau, 2005a)? must try to avoid influencing the participants be- Two key concepts in survey research are pop- ing studied, because even the mere presence of a ulation and sample. A population consists of all the human observer may disrupt a person’s or ani- individuals who we are interested in drawing a conclu- mal’s behavior. Researchers may disguise their sion about, such as “American adolescents.” Be- presence so that participants are not aware of be- cause it is often impractical to study the entire ing observed. Fortunately, when disguise is not population, we would administer the survey to a feasible, people and other animals typically sample, which is a subset of individuals drawn from adapt to and ignore the presence of an observer the larger population. as time passes. This process is called habituation, To draw valid conclusions about a popula- and researchers may delay their data collection tion from the results of a single survey, the sam- until participants have had time to habituate to ple must be representative: A representative the observers’ presence. sample is one that reflects the important character- istics of the population (Figure 2.10). A sample composed of 80 percent males would not be rep- Survey Research: Adolescents’ Exposure resentative of the student body at a college to Abuse and Violence where only 50 percent of the students are men.  Focus 8 In survey research, information about a topic is ob- To obtain a representative sample, survey re- What is random sampling and tained by administering questionnaires or interviews searchers typically use a procedure called why do survey researchers use to many people. Political polls are a well-known ex- random sampling, in which every member of the it? What problems can occur when conducting surveys? ample, but surveys also ask about participants’ population has an equal probability of being chosen behaviors, experiences, and attitudes on wide- to participate in the survey. A common variation of ranging and sometimes sensitive issues. For ex- this procedure, called stratified random sampling, ample, in a recent, carefully conducted national is to divide the population into subgroups based survey of 12- to 17-year-old Americans, 40 percent on characteristics such as gender or ethnic iden- of these adolescents reported that they had wit- tity. Suppose the population is 55 percent fe- nessed violence either at home or in their commu- male. In this case, 55 percent of the spaces in the nity, and 8 percent and 23 percent, respectively, sample would be allocated to women and 45 indicated that they had personally been the vic- percent to men. Random sampling is then used tims of sexual and physical assault (Hanson et al., to select the individual women and men who 2006). This survey studied 3,097 adolescents, who will be in the survey. were interviewed at length by telephone. How is When a representative sample is surveyed, it possible to make an accurate estimate of the we can be confident (though never completely FIGURE 2.10 E Surveys and sampling. A repre- Unrepresentative sentative sample possesses the sample D important characteristics of the Population A population in the same proportions. C Data from a representative sample are more likely to generalize to the E B A larger population than are data from an unrepresentative sample. Sampling D procedure B E A C D B Representative sample C pas82760_ch02_027-059 8/10/07 12:20 PM Page 41 Techbooks (PPG-Quark) STUDYING BEHAVIOR SCIENTIFICALLY 41 certain) that the findings closely portray the pop- time, but it does happen. Thus, for several rea- ulation as a whole. This is the strongest advantage sons, even well-crafted surveys can yield inaccu- of survey research. Modern political opinion polls rate estimates. typically use such excellent sampling procedures that, just prior to elections, they can reasonably predict who will win a national election from a What Do You Think? sample of about 1,000 people. SHOULD YOU TRUST INTERNET AND POP MEDIA SURVEYS? In contrast, unrepresentative samples can Tom fills out a political-attitude survey posted on the produce distorted results. It is better to have a Internet. Claire mails in a dating-satisfaction survey smaller representative sample than a larger, un- that came in a fashion magazine to which she sub- representative one. In one famous example, a mail scribes. Sam responds to a local TV news phone-in survey of almost 2 million voters in 1936 by survey on a tax issue (“Call our number, press ‘1’ to Literary Digest magazine predicted that Republi- agree, ‘2’ to disagree”). For each survey, can the re- can presidential candidate Alf Landon would easily sults be trusted to reflect the general public’s atti- defeat Democratic candidate Franklin Roosevelt. tudes? Think about it, then see page 59. When the election took place, Roosevelt won in a landslide! How could a prediction based on 2 million people be so massively wrong? The answer is that CORRELATIONAL RESEARCH: MEASURING the sample selected for the poll was unrepresenta- ASSOCIATIONS BETWEEN EVENTS tive of the population that voted in the election. What factors distinguish happily married couples  Focus 9 The researchers obtained names and addresses Describe three components of from those headed for divorce? Do firstborn chil- from telephone directories, automobile registra- correlational research and how dren differ in personality from later-born chil- tion lists, and magazine subscription lists. In 1936, they are illustrated by the study dren? Is monetary wealth related to happiness? of very happy people. most poorer Americans did not have telephones, These and countless other psychological ques- cars, or magazine subscriptions. Thus, the sample tions ask about associations between naturally oc- underrepresented poorer socioeconomic groups curring events or variables. To examine such rela- and overrepresented wealthier people: bad sam- tionships, scientists typically conduct correlational ple, bad prediction. In sum, always consider the research, which in its simplest form has three nature of the sample when interpreting survey components: results. In scientific research, surveys are an efficient 1. The researcher measures one variable (X), method for collecting a large amount of informa- such as people’s birth order. tion about people’s opinions, experiences, and 2. The researcher measures a second variable (Y), lifestyles, and they can reveal changes in people’s such as a personality trait. beliefs and habits over many years. But there also 3. The researcher statistically determines whether are several major drawbacks to surveys. First, sur- X and Y are related. vey data cannot be used to draw conclusions about cause and effect. Second, surveys rely on Keep in mind that correlational research in- participants’ self-reports, which can be distorted volves measuring variables, not manipulating by factors such as social desirability bias, a ten- them. dency to respond or behave in a way that is per- Naturalistic observation and surveys are ceived as socially acceptable, rather than respond methods frequently used not only to describe as one truly feels. Participants’ survey responses events but also to study associations between also can be distorted by interviewer bias, by inac- variables. For example, in the naturalistic obser- curate perceptions of their own or other people’s vation study of schoolyard bullying, the re- behavior, and by misinterpreting the survey ques- searchers extensively examined associations be- tions. Third, unrepresentative samples can lead to tween the children’s sex and peer intervention faulty generalizations about how an entire popu- (Hawkins et al., 2001). They found that girls lation would respond. And finally, even when were more likely to intervene when the bully surveys use proper random sampling procedures, and victim were female, and boys were more once in a while—simply by chance—a sample that likely to intervene when the bully and victim is randomly chosen will turn out not to be repre- were male. sentative of the larger population. Overall, in Other types of studies also fall under the cor- properly conducted professional and scientific relational umbrella, as our “Research Close-Up” surveys, this happens less than 5 percent of the illustrates. pas82760_ch02.qxd 8/7/09 5:17 PM Page 42 42 CHAPTER 2 Research Close-Up