Asking and Answering Sociological Questions PDF

Asking and Answering Sociological Questions 2 Sociology today: a relies increasingly on statistical studies that use big data sets. b has no place for empathy. c relies too much on qualitative methods to be a science....

Asking and Answering Sociological Questions 2 Sociology today: a relies increasingly on statistical studies that use big data sets. b has no place for empathy. c relies too much on qualitative methods to be a science. Turn the page for the correct answer. 31  W hen was the last time you posted on a friend’s Facebook page, liked a photo, or changed your relationship status? Many of us use social networking to keep up with friends and family and share details about our lives with a wider audience. Indeed, Facebook not only allows us to keep up with those we care about, but it also serves as a detailed record of our social relationships. In recent years, scholars have taken data from our Facebook pages to ask questions about who we interact with, who we befriend, and who we love. One study claimed that Facebook could predict whether our romantic relationships would last. To make this argument, they drew on a large data set of more than 1 million individuals in romantic relationships, as well as their Facebook friends—a total of nearly 380 million Facebook users. They found that if two people were romantic partners, and their friendship groups overlapped a great deal, meaning they had the same group of friends, they had a higher likelihood of staying together (Backstrom and Kleinberg, 2013). If your answer to the question on the previous page was that sociology today relies more and more on studies that, like this one, use big data, then you are correct. Increas- ingly, studies are relying on statistical or quantitative methods and computer pro- grams that make it possible to analyze the vast amounts of data the Internet generates. These methods allow sociologists to process more data than ever before. Even as we reap all the scientific potential +/(.#..#0ŗ'.")-ŚĊŚApproaches to that can be seen in the age of big data, we must sociological research that draw on objective and also contend with its potential arrogance. statistical data and often focus on documenting trends, comparing subgroups, or exploring Some researchers feel more confident than correlations. ever before in the claims that can be made from big data sets. Researchers who have LEA R NING OBJECTIV ES 1 BASIC CONCEPTS Learn the steps of the research process. Name the different types of questions sociologists address in their research—factual, theoretical, comparative, and developmental. 2 ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS: HISTORICAL CONTEXT Contrast Park’s and Ogburn’s visions of sociology as a science. Understand their influence on contemporary sociological research. 3 ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS TODAY: RESEARCH METHODS Familiarize yourself with the methods available to sociological researchers and recognize the advantages and disadvantages of each. 4 UNANSWERED QUESTIONS Understand how research methods generate controversies and ethical dilemmas for sociologists. 32 CHAPTER 2 Asking and Answering Sociological Questions never been to a slum, nor have any personal relationship to a single poor person, feel they +/&#..#0ŗ'.")-ŚĊŚApproaches to sociological research that often rely on personal can speak with authority about inequality and/or collective inter views, accounts, or and poverty from a study based on interesting observations of a person or situation. statistical correlations. In the age of big data, empathy and involvement with communities and people under study are less frequently the basis of social scientific insights. But if you answered b— sociology has no place for empathy—that would not be correct either. Sociology has a rich tradition that also includes qualitative methods, which rely on observations, interviews, and archival data. Here, personal involvement and empathy count for a great deal, though most of the insights in qualitative studies also derive from other kinds of sociological thinking as well. While quantitative analyses may make use of numer- ical data (such as those collected in surveys or Facebook friend lists), qualitative analyses may use data derived from interactions, interviews, conversations, or observations of a social scene. For example, a qualitative study of the same question asked in the Facebook study of romantic relationships might use fewer cases to focus on richer details of how individuals dissolve their relationships. In Uncoupling, Diane Vaughan (1986) explored this question by using interviews with people who broke up to better understand how relationships end. Does the reliance on qualitative methods make sociology less of a science? It does not. Do quantitative methods make sociology more scientific? They do not. When many people think of a “science,” they typically imagine fields like physics or chemistry. Yet, as we shall see, what makes chemistry and physics a science is not the sample size or even the subject matter, but rather a set of values that can be deployed in any research. In this chapter, we begin by discussing those values to clarify the meaning of science in social science. T HE ANSWER I S A. 1 BASIC CONCEPTS Regardless of whether it is quantitative or qualitative, sociological research that is striving to be scientific tries to meet basic standards (King, Keohane, and Verba, 1994). First, the goal of sociological research is inference. By this we mean that when we make observations particular to a specific setting or group, the goal is to be able to generalize beyond that specific entity to others of its kind. In other words, while sociologists can- not collect data about the whole world, they are able to use more limited data to make broader claims about phenomena they cannot directly observe (King et al., 1994). Take as an example Uncoupling. If Diane Vaughan limited her claim to the 103 interviews she recorded, this material on its own would not constitute a sociological analysis. Moving beyond these interviews to make more general claims about how we—as humans—form and break romantic relationships is part of what makes her analysis scientific. Second, sociologists must ensure that other researchers can retrace the paths to their findings. Ideally, others in the scientific community can reproduce their results. The procedures researchers use for collecting and analyzing data should be public; the reliability of research data can only be verified if the ways in which that data was Basic Concepts33  collected and processed are made explicit to the scientific community, allowing other researchers both to learn from the methods used and to address the limitations of these methods. What’s more, careful documentation of the way one arrives at conclusions about the social world allows research findings to be comparable even when conducted by researchers at different times. Third, the conclusions of all scientific research, including sociology, are uncertain. This might be surprising. Isn’t the purpose of studying the social world to be able to make force- ful arguments about it? The values of science sometimes demand that we do the opposite. The scientific validity of inferences can be assessed only if researchers are clear about all the sources of their uncertainty. In sociology, like any science, the highest status is con- ferred on those who are honest about how certain their conclusions really are. Sociologists need to specify all the possible sources of uncertainty in their study. All three of the previous standards of empirical research are found in both the natural sciences and the social sciences. They are also shared by quantitative and qualitative social science. It is important to highlight a fourth element that is emphasized more in the social sciences—particularly in qualitative research. This is called reflexivity. For social scientists, it is particularly important to acknowledge that the investigator is a crucial part of the world she studies and cannot necessarily divorce herself from it. This includes power dynamics among subjects and the ways in which personal val- ues or personal identity influence both the nature of the questions asked and the inter- pretation of data. Social scientists must thus be ready to reflect on how the way they are part of the social scenes they study may affect the kinds of conclusions they draw. For example, a middle- class researcher studying a poor population should be clear about how his or her class position influenced a sociological argument or the relations with subjects. These four principles help identify work that is living up to the highest ideals of social science. Some good scientific work might not necessarily be strong in all the areas at once, but the goal of social science should be to achieve as many of them as possible. While these principles are useful for understanding how sociology can strive to be a sci- ence, they don’t provide a practical guide for how one might begin defining and going on to answer a research question. To do this, we now consider the research process. The Research Process In order to understand the way sociology asks and answers questions, it is helpful to think of the research it does as a process. We can better understand the main concepts of research design by breaking down the process into seven stages of research—beginning with the definition of a research question and ending with the dissemination of the study findings. 1. DEFINE THE RESEARCH PROBLEM All research starts with a research problem. This problem may be an area of factual igno- rance about, say, certain institutions, social processes, or cultures. A researcher might seek to answer questions such as, What proportion of the population holds strong reli- gious beliefs? Are people today disaffected with “big government”? How far does the eco- nomic position of women lag behind that of men? Do LGBTQ and straight teens differ in their levels of self-esteem? 34 CHAPTER 2 Asking and Answering Sociological Questions  The best sociological research begins with problems that are also puzzles. Figure 2.1 A puzzle arises not simply from a lack Steps in the Research Process of information but also from a gap in our understanding. Much of the skill in producing worthwhile sociological DEFINE THE PROBLEM research consists in correctly identify- Select a topic for research. ing puzzles. Rather than just answering the REVIEW THE LITERATURE question, What is going on here? skilled Familiarize yourself with existing researchers try to illuminate why events research on the topic. happen as they do. Thus, we might ask, Why are patterns of religious belief FORMULATE A HYPOTHESIS changing? What accounts for the recent What do you intend to test? What is the decline in the proportion of the popula- relationship among the variables? tion voting in presidential elections? Why are women poorly represented in science and technology jobs? What are SELECT A RESEARCH DESIGN Choose one or more research methods: the characteristics of high schools with experiment, survey, observation, high levels of bullying? use of existing sources. No piece of research stands alone. One project may lead to another because CARRY OUT THE RESEARCH it raises issues the researcher had not Collect your data; record information. previously considered. A sociologist may discover puzzles by reading the INTERPRET YOUR RESULTS work of other researchers in books Work out the implications of the data and professional journals or by being you collect. aware of social trends. For example, an increasing number of public health REPORT THE RESEARCH FINDINGS programs have sought to treat the men- What is their significance? How do they tally ill while they continue to live in relate to previous findings? the community rather than confining them in asylums. Sociologists might be prompted to ask, What has caused this Your findings are registered and discussed in the wider academic community, leading shift in attitude toward the mentally ill? perhaps to the initiation of further research. What are the likely consequences for the patients themselves and for the rest of the community? 2. REVIEW THE LITERATURE Once a research problem is identified, the sociologist must review related research: Have previous researchers spotted the same puzzle? How have they tried to solve it? What aspects of the problem has their research left unanalyzed? Have they looked only at small segments of the population, such as one age group, gender, or region? Drawing on oth- ers’ ideas helps the sociologist clarify the relevant issues and the appropriate research methods. Basic Concepts35  3. MAKE THE PROBLEM PRECISE A third stage involves clearly formulating the research problem. If relevant literature already exists, the researcher may have a good idea of how to approach the problem. At this stage, hunches sometimes become hypotheses — educated guesses about what is going on. For the research to be effective, the researcher must formulate a hypothesis in such a way that the factual material gathered will provide evidence either supporting or disproving it. 4. WORK OUT A DESIGN The researcher then decides how to collect the research materials, choosing from a range of methods based on the study objectives as well as the aspects of behavior under study. For some purposes, a survey (usually involving questionnaires) might be suitable. In other circumstances, interviews or an observational study might be appropriate. 5. CARRY OUT THE RESEARCH Researchers then proceed to carry out the plan developed in step 4. However, during the actual research, unforeseen practical difficulties may arise that force the researcher to rethink his or her initial strategy. For example, it might prove impossible to contact certain questionnaire recipients or interview subjects. A business firm or government agency might not let the researcher carry out the work as planned. Yet omitting such persons or institutions from the study could bias the results, creating an inaccurate or incomplete picture of social reality. For instance, a researcher studying how business corporations have complied with affirmative action programs might find that companies that have not complied do not wish to be studied. 6. INTERPRET THE RESULTS Once the information has been gathered, the researcher’s work is not over—it is just beginning! The researcher must analyze the data, track trends, and test hypotheses. Most important, researchers must interpret their results in such a way that they tell a clear story and directly address the research puzzle outlined in step 1. Although it may be pos- sible to reach clear answers to the initial questions, many investigations are ultimately not fully conclusive. 7. REPORT THE FINDINGS The research report, usually published as a journal article or book, provides an account of the research question, methods, findings, and the implications of the findings for social theory, public policy, or practice. This stage is only final in terms of the individual project. Most reports identify unanswered questions and suggest new questions for further research. All individual research investigations are part of the continuing process of research within the sociological community. "3*)."--ŚĊŚIdeas or educated guesses about REALITY INTRUDES! a given state of affairs, put forward as bases for empirical testing. The preceding sequence of steps is a simpli- fied version of what happens in actual research 36 CHAPTER 2 Asking and Answering Sociological Questions projects. These stages rarely succeed each other so neatly; the difference is like that between the recipes outlined in a cook- book and the actual process of preparing a meal. Experienced cooks often don’t work from recipes at all, yet they might cook better meals than those who do. CONCEPT CHECKS 3 1. Compare and contrast qualitative and quanti- Following fixed schemes can be unduly tative research methods. restricting; much outstanding sociologi- 2. What are the seven steps of the research cal research would not fit rigidly into this process? sequence, though it would include most of 3. What is a hypothesis? the steps outlined here. 2 ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS: HISTORICAL CONTEXT When sociology began as a discipline, it was a highly theoretical field. It consisted of much armchair speculation, and many of the notions it developed about how the world worked were not well grounded in evidence. But in the 1920s, there developed in Ameri- can sociology, largely at the University of Chicago, a more intense commitment to the idea that such theoretical speculations were not enough—that sociology as a discipline needed to ground its concepts and theories in facts and data. This goal for sociology was represented in two figures, both of whom were professors at the University of Chicago: Robert Park (1864–1944) and William Ogburn (1886–1959). Park’s beliefs about how to make social research more scientific came from his background, both as a student of philosophy in Europe and as a reporter for the Minneapolis Star. He was interested in developing theories but wanted them to relate directly to the actual lives of people and to be based on the careful accumulation of evidence about people’s lives. He told his students that to do real research they needed to get the seat of their pants dirty, to wear out their shoe leather to discover the truth. Park thought that the most impor- tant thing for a sociologist to do was to go around all the neighborhoods of the (left) Robert Park (1864–1944) and (right) William Ogburn (1886–1959) city and find out what was going on by meeting the people who were the sub- jects of the sociologists’ theories. Fol- lowing Park’s lead, the University of Chicago’s sociology department used the city as a laboratory. Its sociologists took on roles in the community to see how the community’s members lived, conducting interviews and firsthand observation. Their research reports tended to be highly systematic, well written, and oriented toward improv- ing conditions in the city and around the United States. Asking and Answering Sociological Questions: Historical Context37  William Ogburn, however, Park’s colleague in the department of sociology at the University of Chicago, didn’t believe that the future of sociology lay in shoe leather, well- written books, findings that could not be quantified, or efforts to influence public policy. These, he thought, were the domain of ethics, religion, journalism, and propaganda. In his presidential address to the American Sociological Society, he argued that sociology needed to become a science (Ogburn, 1930). The goal, he argued, was not “to make the world a better place in which to live” or to set forth “impressions of life” or to “[guide] the ship of state” but only to “[discover] new knowledge.” Ogburn wanted sociology to be a field that looked a lot more like the natural sciences in both its presentation and its orienta- tion. Whereas Park had clear ideas about what the subject matter of sociology should be (immigration and the life of the city), Ogburn believed sociologists could study anything that could be measured with numbers. These two figures, Park and Ogburn, coexisted at the University of Chicago for many years, each committed to his own vision of sociology as a science. For Park and his stu- dents, the personal, emotional, and scientific side of sociology complemented the aspira- tion to develop explanations about the social world. We can see this legacy in the work of young sociologists today. But the legacy of William Ogburn is no less significant. The value of statistics and scientific method- ologies for understanding the world has never been greater, in part because of the CONCEPT CHECK 3 massive amount of data and information being collected on the Internet. More than ever before, business and government need 1. How did Park and Ogburn approach socio- logical research differently? In what ways did people who can analyze this material and each sociologist influence researchers today? who are disposed to think about the world in a scientific way. 3 ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS TODAY: RESEARCH METHODS Let’s look at the various research methods sociologists currently employ (Table 2.1). Whereas many fields have one main method, modern sociology embraces a variety of methodologies. People trained in sociology end up developing a wide range of researchs- kills, which makes them quite versatile after graduation. Here are the basic methods used ,-,"ŗ'.")-ŗŚĊŚThe diverse methods of investigation used to gather empirical (factual) today, with examples drawn from recent material. Different research methods exist in research. sociology, but the most commonly used are fieldwork (or participant observation) and survey methods. For many purposes, it is useful to Ethnography combine two or more methods within a single research project. One widely used qualitative method is ."()!,*"3ŚĊŚThe firsthand study of people using ethnography, or firsthand studies of people observation, in-depth interviewing, or both. Also using observations, interviews, or both. Here, called fieldwork. the investigator socializes, works, or lives with participant observation Ċ A method of research members of a group, organization, or commu- widely used in sociology and anthropology in which the researcher takes part in the activities of the nity. In the case of participant observation, group or community being studied. the researcher participates directly in the activ- ities he or she is studying. Other ethnographers, 38 CHAPTER 2 Asking and Answering Sociological Questions  Table 2.1 Three of the Main Methods Used in Sociological Research RESEARCH METHOD STRENGTHS LIMITATIONS Ethnography Usually generates richer and more Can be used to study only in- depth information than other relatively small groups methods. or communities. Can provide a broader Findings might apply only to understanding of social processes. groups or communities studied; not easy to generalize on the basis of a single fieldwork study. Surveys Make possible the efficient Material gathered may be collection of data on large numbers superficial; if questionnaire is of individuals. highly standardized, important Allow for precise comparisons to differences among respondents’ be made among the answers of viewpoints may be glossed over. respondents. Responses may be what people profess to believe rather than what they actually believe. Experiments Influence of specific variables can Many aspects of social life cannot be controlled by the investigator. be brought into the laboratory. Are usually easier for subsequent Responses of those studied may researchers to repeat. be affected by the experimental situation. by contrast, may observe at a distance and not participate directly in the activities under observation. An ethnographer cannot simply be present in the group she studies, but must explain and justify her presence to its members. The researcher must gain and sustain the cooperation of the community to achieve worthwhile results. For a long while, research based on participant observation excluded accounts of the hazards or problems involved, but more recently, field-workers have been more open. Frequently, field-workers experience feelings of loneliness and frustration, the latter occurring especially when group members refuse to talk frankly with them. Some types of fieldwork may be physically dangerous; for instance, a researcher studying a delin- quent gang might be seen as a police informer or might become unwittingly embroiled in conflicts with rival gangs. In traditional works of ethnography, accounts provided little information about the observer because ethnographers were expected to present objective reports. More recently, ethnographers have increasingly spoken about their connection to the people under study. For example, a researcher might discuss how her race, class, or gender affected the work, or how the power differences between observer and observed dis- torted the dialogue between them. ADVANTAGES AND LIMITATIONS OF FIELDWORK Successful ethnography provides rich information on the behavior of people in groups, organizations, and communities, as well as information on how these people understand Asking and Answering Sociological Questions Today: Research Methods39  their own behavior. Once we look inside a given group, we can better understand not only that group but also broader social processes. But fieldwork has its limitations. Only fairly small groups or communities can be studied. Much also depends on the researcher’s skill in gaining the confidence of the individuals involved. Also, a researcher could identify so closely with the group that he loses the perspective of an objective outside observer. Or a researcher might draw conclusions that are more about his own effect on the situation than he or his readers real- ize. Interpreting ethnographies usually involves Sociologist Alice Goffman spent six years doing par- ticipant observation research in a poor neighborhood problems of generalizability because we cannot of Philadelphia for her study of the impact of intensive be sure that what we find in one context will apply policing on young black men. in others, or even that two different researchers will draw the same conclusions when studying the same group. A RECENT EXAMPLE When Alice Goffman was a graduate student, she spent six years in an intensely policed poor black neighborhood. She was interested in the social situation of large numbers of black men who were on the run from the criminal justice system. She became part of the everyday life of a group of boys who were known as the 6th Street Boys, writing about them in her book On the Run (Goffman, 2014). Goffman’s research cast new light on the struggles of men who were dipping and dodging the police, worrying that any encounter would result in their imprisonment. By spending time with them every day for many years, Goffman was able to see that for these men, activities, relations, and localities that others relied on to maintain a decent and respectable identity were trans- formed into a system that authorities used to locate, arrest, and confine them. The police and the courts became dangerous to interact with, as did showing up to work or going to places such as hospitals. Instead of a safe place to sleep, eat, and find acceptance and support, their childhood homes were transformed into a “last known address,” one of the first places police looked for them. Close relatives, friends, and neighbors became poten- tial informants. Surveys Quantitative methodologists have a range of analytical tools and data resources at their disposal, but surveys are the most commonly used. When conducting a survey, researchers ask subjects to provide answers to structured questionnaires, which are administered in person or mailed to a select group of people. This group is survey Ċ A method of sociological research known as a population. While ethnogra- in which questionnaires are administered to the population being studied. phies are best suited for in- depth studies of population Ċ The people who are the focus of small slices of social life, survey research pro- social research. duces information that is less detailed but can be generalized to the population as a whole. 40 CHAPTER 2 Asking and Answering Sociological Questions STANDARDIZED AND OPEN-ENDED QUESTIONS *#&).ŗ-./3ŚĊŚA trial run in survey research. -'*&ŚĊŚA small proportion of a larger Two types of questions are used in surveys. population. With standardized, or fixed-choice, questions, ,*,-(..#0ŗ-'*&ŚĊŚA sample from a only a fixed range of responses is possible—for larger population that is statistically typical of example, Yes, No, Don’t know or Very likely, that population. Likely, Unlikely, Very unlikely. Such questions -'*&#(!ŚĊŚStudying a proportion of individuals or cases from a larger population as have the advantage that responses are easy to representative of that population as a whole. count and compare because only a small num- ,()'ŗ-'*&#(!ŚĊŚSampling method in which ber of categories are involved. However, because a sample is chosen so that every member of the they do not allow for subtleties of opinion or ver- population has the same probability of being included. bal expression, they may yield restrictive, if not misleading, information. Open- ended questions, by contrast, typically provide more detailed information because respondents may express their views in their own words, and the researcher can ask follow-up questions to probe more deeply into what the respondent thinks. However, the lack of standardization means that responses may be difficult to compare statistically. In surveys, all the items must be readily understandable to interviewers and interviewees alike. Questions are usually asked in a set order. In large national surveys undertaken by government agencies and research organizations, interviews occur more or less simultaneously across the country. Those who conduct the interviews and those who analyze the results could not work effectively if they constantly had to check with one another about ambiguities in the questions or answers. Questionnaires should also accommodate the characteristics of respondents. Will they see the point of a particular question? Might it offend them? Do they have enough information to answer usefully? Will they answer at all? A questionnaire’s terminology might be unfamiliar; for instance, “What is your marital status?” might better be asked as, “Are you single, married, separated, or divorced?” Most surveys are preceded by pilot studies, which reveal problems with the survey not anticipated by the investigator. A pilot study is a trial run in which just a few people participate. Any difficulties can then be ironed out before the main survey takes place. SAMPLING Often sociologists are interested in the characteristics of large numbers of individuals— for example, political attitudes of the American population as a whole. In such situations, researchers concentrate on a sample, or a small proportion of the overall group. Usu- ally, the results from a properly chosen sample can be generalized to the total population. Studies of only 2,000–3,000 voters, for instance, can accurately indicate the attitudes and voting intentions of the entire population. But to achieve such accuracy, we need a representative sample; the group of individuals studied must be typical, or representa- tive, of the population as a whole. Because sampling is highly complex, statisticians have developed rules for working out the correct size and nature of samples. A particularly important procedure that ensures that a sample is representative is random sampling , in which every member of the sample population has the same probability of being included. The most sophisticated way of obtaining a random sample is to assign each member of the population a number and then use a computer to gener- ate a random list from which the sample is derived—for instance, by picking every tenth Asking and Answering Sociological Questions Today: Research Methods41  STATISTICAL TERMS Research in sociology often makes use of statisti- In such instances, one of two other measures cal techniques in the analysis of findings. Some are may be used. The mode is the figure that occurs most highly sophisticated and complex, but those most often frequently in a given set of data. In our example, it is used are easy to understand. The most common are $40,000. The problem with the mode is that it doesn’t measures of central tendency (ways of calculating take into account the overall distribution of the data— averages) and correlation coefficients (measures of that is, the range of figures covered. The most frequently the degree to which one variable relates consistently occurring case in a set of figures is not necessarily rep- to another). resentative of the distribution as a whole and thus may There are three methods of calculating aver- not be a useful average. In this example, $40,000 is too ages, each of which has certain advantages and short- close to the lower end of the figures. comings. Take as an example the amount of personal The third measure is the median, which is the wealth (including all assets, such as houses, cars, bank middle of any set of figures; here, this would be the sev- accounts, and investments) owned by 13 individuals. enth figure, again, $40,000. Our sample includes an odd Suppose the 13 own the following amounts: number of figures, 13. If there were an even number— for instance, 12—the median would be calculated by taking the mean of the two middle cases, figures 6 and 7. 1 $0 8 $80,000 As with the mode, the median gives no indication of the actual range of the data being measured. 2 $5,000 9 $100,000 Sometimes a researcher will use more than one 3 $10,000 10 $150,000 measure of central tendency to avoid giving a decep- 4 $20,000 11 $200,000 tive picture of the average. More often, a researcher 5 $40,000 12 $400,000 will calculate the standard deviation for the data 6 $40,000 13 $10,000,000 in question. This is a way of calculating the degree of 7 $40,000 dispersal, or the range, of a set of figures—which in this case goes from $0 to $10,000,000. Correlation coefficients offer a useful way of The mean corresponds to the average, arrived at expressing how closely connected two (or more) by adding together the personal wealth of all the peo- variables are. When two variables correlate com- ple and dividing the result by the number of people in pletely, we can speak of a perfect positive correlation, the sample (13). The total is $11,085,000; dividing this expressed as 1.0. When no relation is found between amount by 13, we calculate the mean to be $852,692.31. two variables—they have no consistent connection at The mean is often a useful calculation because it is all—the coefficient is 0.0. A perfect negative correla- based on the whole range of data provided. However, tion, expressed as 21.0, exists when two variables are the mean can be misleading when one or a small num- in a completely inverse relation to one another. Perfect ber of cases is very different from the majority. In this correlations are never found in the social sciences. example, the mean is not in fact an appropriate measure Correlations of the order of 0.6 or more, whether posi- of central tendency, because the presence of one very tive or negative, are usually regarded as indicating a large figure, $10,000,000, skews the picture. One might strong degree of connection between whatever vari- get the impression, when using the mean to summarize ables are being analyzed. Positive correlations on this these data, that most of the people own far more than level might be found between, say, social class back- they actually do. ground and voting behavior. number. Random sampling is often done by researchers doing large population-based surveys, aimed to capture the behaviors or attitudes of the overall U.S. population. For qualitative researchers interested in a particular population, such as street vendors or gangsters, it simply would not make sense to try to draw a random sample. ADVANTAGES AND DISADVANTAGES OF SURVEYS Surveys are widely used in sociological research for several reasons. Questionnaire responses can be more easily quantified and analyzed than material generated by most 42 CHAPTER 2 Asking and Answering Sociological Questions other research methods; large numbers of peo- ple can be studied; and, given sufficient funds, '-/,-ŗ) ŗ(.,&ŗ.((3ŚĊŚThe ways of calculating averages. researchers can employ a specialized agency to ),,&.#)(ŗ)ö#(.-ŚĊŚThe measure of the collect the responses. The scientific method is degree of correlation between variables. the model for this kind of research since surveys '(ŚĊŚA statistical measure of central give researchers a statistical measure of what tendency, or average, based on dividing a total by they are studying. the number of individual cases. However, many sociologists are critical of the ')ŚĊŚThe number that appears most often survey method. They argue that findings whose in a given set of data. This can sometimes be a helpful way of portraying central tendency. accuracy may be dubious—given the relatively '#(ŚĊŚThe number that falls halfway in a shallow nature of most survey responses—can range of numbers—a way of calculating central nonetheless appear to be precise. Also, levels tendency that is sometimes more useful than of nonresponse are sometimes high, especially calculating a mean. now that so many people use cell phones and no -.(,ŗ0#.#)(ŚĊŚA way of calculating the spread of a group of numbers. longer have landlines at home. Furthermore, some published studies are based on results !,ŗ) ŗ#-*,-&ŚĊŚThe range or distribution of a set of figures. derived from little more than half a sample, though normally there is an effort to recontact nonrespondents or to substitute them with other people. Although little is known about those who do not respond to surveys or who refuse to be interviewed, we do know that people often experience survey research as intrusive and time-consuming. One devel- opment that may hold great promise for learning about public opinion is the rise of new statistical techniques for surveying the online conversation that is taking place quite naturally on the Internet. A RECENT EXAMPLE One of the most famous contemporary surveys is the General Social Survey, which has been administered to Americans since 1972. It is sometimes called the “pulse of America,” as it has tracked the social life of Americans for decades. Since 1985, however, it has also been administered outside the United States to obtain comparative data. One of its most significant and controversial findings of recent years came from an analysis of the num- ber of real-life “friends” Americans were reporting. The results suggested that Americans had fewer confidants than in the past, and that a growing number couldn’t name a single person with whom they shared “important matters.” The implication was that Americans were growing lonelier (McPherson, Smith-Lovin, and Brashears, 2006). How did the researchers evaluate whether friendships were declining over time? What accounts for this decline? The study was based on face-to-face interviews with a nationally representative sample of nearly 1,500 American adults. All had participated in the long-running General Social Survey and were asked questions about their social networks. Specifically, they were asked to identify people with whom they had discussed “matters [that are] important to [them]” in the past six months. On average, they named 2.08 people in 2004, compared to 2.94 people in 1985. The proportion of respondents who reported that there was no one with whom they discussed important matters jumped from 10 percent in 1985 to 25 percent in 2004. Some social scientists are not convinced that these findings support the claim that Americans are isolated or lonely. Rather, some argue that “weak” social ties, such as those with acquaintances, may be perfectly acceptable and rewarding for some people; these people may actually prefer to have many casual acquaintances rather than a handful of Asking and Answering Sociological Questions Today: Research Methods43  deep friendships. University of Toronto sociolo- gist Barry Wellman (1994), for example, believes that the study offers important findings about “intimate ties,” but he questions whether these findings should be taken as evidence that Amer- icans are lonely and isolated. Rather, he notes that people’s overall ties are actually increasing compared to previous decades, due in part to the Internet. Experiments Experiments are often used in the natural sci- ences and psychology, as they are considered the best method for ascertaining causality, or the influence of a particular factor on a study’s outcome. In an experimental situation, the researcher directly controls the circumstances being studied. In a typical experiment, people are randomly assigned to two groups. The first, called an experimental group, receives some spe- cial attention based on the researcher’s theory; the second, the control group, does not receive this attention. The subjects usually do not know From his jail experiment, Zimbardo concluded that to which group they have been assigned and sel- behavior in prisons is influenced more by the nature of dom know the purpose of the experiment, though the prison itself than by the individual characteristics of this is not always the case. those involved. A classic example is the 1971 experiment carried out by Philip Zimbardo (1992), who set up a make-believe jail, randomly assigning some student volunteers to the role of guard and other volunteers to the role of prisoner. His aim was to see how role-playing would affect changes in attitude and behavior. The results shocked the investigators. Students who played guards quickly assumed an authoritarian manner; they displayed real hostility toward the prisoners, verbally abusing and bullying them. The prisoners, by contrast, showed a mixture of apathy and rebelliousness—a response often noted among inmates in actual prisons. These effects were so marked and the level of tension so high that the experiment had to be called off at an early stage. The results, however, were important: Zimbardo concluded that behavior in prisons is more influenced by the nature of the prison situation than by the individual characteristics of those involved. ADVANTAGES AND DISADVANTAGES OF EXPERIMENTS The advantage of experimental studies is that researchers can test a hypothesis under highly controlled conditions established by the researcher. The ability to con- trol experimental conditions, however, is also experiment Ċ A research method by which the principal weakness of experimental stud- variables can be analyzed in a controlled and ies, which in many ways are artificial. To the systematic way, either in an artificial situation constructed by the researcher or in a naturally extent that the laboratory fails to duplicate occurring setting. a natural setting, it is difficult to generalize the results of laboratory experiments to the 44 CHAPTER 2 Asking and Answering Sociological Questions larger society. We can bring only small groups of individuals into a laboratory setting, and, in comparative research Ċ Research that compares one set of findings on one society with such experiments, people know they are being the same type of findings on other societies. studied and may behave unnaturally. As a result, sociologists sometimes use field experi- ments, in which a real-life situation is simu- lated as accurately as possible. A RECENT EXAMPLE In 1994, a group of sociologists launched an experiment to find out if it makes a differ- ence to move people from high-poverty ghetto neighborhoods to low-poverty neighbor- hoods. People who responded to an advertisement offering vouchers for new apartments were randomly assigned to two groups—those who were given the opportunity to move (experimental group) and those who weren’t (control group). Why did the experimenters compare the fates of two groups who both applied for vouchers, one of which got them and one of which didn’t? Why not compare people who got the vouchers against those who did not apply in the first place? In fact, the study kept data on all three groups. The ones who applied but did not receive vouchers had more in common with the ones who got to move, because they also wanted to escape from their current circumstances. The best comparison is between two groups that wanted to move. They are alike in all characteristics except for whether they actually stayed in place or experienced a new living condition. Following the groups since the mid-1990s, researchers made some important discoveries. For example, people who moved did not do much better in the labor market than those who stayed behind. But the movers tended to be much happier, and they experienced decreased levels of obesity when compared with those who did not get to move. Comparative Historical Research Comparative research is of central importance in sociology because it enables researchers to document whether social behavior varies across time and place and by one’s social group membership. Most comparative work is quantitative in that research- ers aim to document whether behaviors and attitudes change over time and place; thus, a consistent metric is required to make comparisons. Consider the American rate of divorce—the number of divorces per thousand married people. Divorce rates rose rap- idly in the United States after World War II, reaching a peak in 1979. Since then, the divorce rate has dropped by nearly a quarter, with only 16.9 marriages per 1,000 end- ing in divorce in 2015 (Anderson, 2016)—a statistic expressing profound changes in the area of sexual relations and family life. Do these changes reflect specific features of American society? We can find out by comparing divorce rates in the United States with those in other countries. The comparison reveals that although the U.S. rate is higher than the rate in most other Western societies, the overall trends are similar. The most influential way of doing comparative research is through historical research. One classic study that investigated a much longer period and applied com- parative research in a historical context was Theda Skocpol’s States and Social Revolu- tions (1979), one of the best-known studies of social change. To produce a theory of the origins and nature of revolution grounded in detailed empirical study, Skocpol looked at processes of revolution in three historical contexts: the 1789 revolution in France, Asking and Answering Sociological Questions Today: Research Methods45  READING A TABLE When reading sociological literature, you will often sometimes gives you some insight into how reliable come across tables. They sometimes look complex but the information is likely to be and tells you where to are easy to decipher if you follow the few basic steps find the original data. In Table 2.2, the source note outlined here; with practice, these will become auto- makes clear that the data have been taken from one matic. (See Table 2.2 as an example.) Do not succumb international organization. to the temptation to skip over tables; they contain 3. Read the headings along both the top and the information in concentrated form that can be read more left-hand side of the table. (Sometimes tables are quickly than would be possible if the same material were arranged with “headings” at the foot rather than expressed in words. By becoming skilled in the interpre- the top.) These tell you what type of information is tation of tables, you will also be able to check how justi- contained in each row and column. In reading the fied the conclusions a writer draws actually are. table, keep in mind each set of headings as you scan the figures. In our example, the headings on the left 1. Read the title in full. Tables frequently have long name the countries involved, while those at the top titles that represent the researcher’s attempt to state refer to the proportion of the population who hold accurately the nature of the information conveyed. a “favorable” opinion of the United States and the The title of Table 2.2 gives, first, the subject of the years for which data are available. data; second, the fact that the table provides material 4. Identify the units used; the figures in the body of the for comparison; and third, the fact that data are given table may represent cases, percentages, averages, only for a limited number of countries. or other measures. Sometimes it may be helpful to 2. Look for explanatory comments, or notes, about convert the figures to a form more useful to you: If the data. A source note at the foot of Table 2.2 indi- percentages are not provided, for example, it may be cates that the data were obtained from the Pew worth calculating them. Research Center, a large international survey orga- 5. Consider the conclusions that might be reached nization. It also notes that data were not available from the information in the table. Most tables are for all nations for all years. Footnotes may say how discussed by the author, and what he or she has to the material was collected or why it is displayed in say should, of course, be considered. But you should a particular way. If the researcher did not gather the also ask what further issues or questions the data data but instead used findings originally reported might suggest. How might you explain some of these elsewhere, a source will be included. The source declines? Or the sudden and precipitous drops? the 1917 revolution in Russia (which brought the Communists to power and established the Soviet Union, which was eventually dissolved in 1989), and the revolution of 1949 in China (which created Communist China). By analyzing a variety of documentary sources, Skocpol was able to develop a power- ful explanation of revolutionary change, one that emphasized underlying social struc- tural conditions. She showed that social revolutions are largely the result of unintended consequences. Before the Russian Revolution, for instance, various political groups were trying to overthrow the regime, but none of these groups—including the Bolsheviks (Communists), who eventually came to power—anticipated the revolution that occurred. A series of clashes and confrontations gave rise to a process of social transformation that was much more radical than anyone had foreseen. At the time that Skocpol wrote, existing theories basically related the emergence of revolutions to the strength of social movements, and these to class relations. Skocpol showed, first, that state structures are as important as class relations and more important than the strength of the revolution- ary movements; and second, that these state structures are heavily influenced by interna- tional events (for instance, revolutions come in the wake of a breakdown in state authority often due to lost international wars). 46 CHAPTER 2 Asking and Answering Sociological Questions  Table 2.2 Opinion of the United States: Comparison of Selected Nations Several interesting trends can be seen from the data in this table. First, the extent to which people hold favorable attitudes toward the United States varies considerably across nations. Second, although there are strong national and regional patterns of support for the United States, we also see considerable historical variation within nations. For example, in Egypt, one-third held favorable views in 2006, yet this proportion has steadily gone downward and bottomed out at 10 percent in 2014. Yet some nations show a steep and sudden drop rather than a steady decline. While two- thirds of Mexicans held a favorable view through much of the 2000s, this share had plummeted to 30 percent by 2017. PERCENTAGE OF PERSONS WHO HOLD A “FAVORABLE” (VS. “UNFAVORABLE”) OPINION OF THE UNITED STATES COUNTRY 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2017 China 42 47 34 41 47 58 44 43 40 50 44 - Egypt - 30 21 22 27 17 20 19 16 10 - - France 43 39 39 42 75 73 75 69 64 75 73 46 Germany 42 37 30 31 64 63 62 52 53 51 50 35 Indonesia 38 30 29 37 63 59 54 - 61 59 62 48 Japan - 63 61 50 59 66 85 72 69 66 68 57 Jordan 21 15 20 19 25 21 13 12 14 12 14 15 Mexico - - 56 47 69 56 52 56 66 63 66 30 Pakistan 23 27 15 19 16 17 12 12 11 14 22 - Poland 62 - 61 68 67 74 70 69 67 73 74 73 Russia 52 43 41 46 44 57 56 52 51 23 15 41 S. Korea - - 58 70 78 79 - - 78 82 84 75 Spain 41 23 34 33 58 61 64 58 62 60 65 31 Turkey 23 12 9 12 14 17 10 15 21 19 29 18 U.K. 55 56 51 53 69 65 61 60 58 66 65 50 U.S. 83 76 80 84 88 85 79 80 81 82 83 - Source: Pew Research Center, 2017g. A RECENT EXAMPLE Such studies as those pioneered by Theda Skocpol rely on qualitative methods of his- torical research, which depend on careful comparisons of a small number of cases. One of the costs of such approaches has been that sociologists have derived their under- standings from some very famous revolutions or wars, while ignoring most of the world altogether. Asking and Answering Sociological Questions Today: Research Methods47  In recent years, Andreas Wimmer has taken up an alternative approach to histori- cal sociology. He uses formal modeling and statistical techniques to analyze hundreds of cases at the same time, rather than a few famous ones. Unlike scholars who go to the library and pull existing data off the shelf, Wimmer has found it necessary to create his own data sets to answer age-old questions. In his approach, what happened in Haiti or in Latin America counts no less than what went on in Holland or in Europe more generally— though these latter places have tended to be studied much more frequently and thus count more heavily in existing theories. Wimmer’s work therefore goes against the grain of most historical sociology that tends to focus mainly on Europe at the expense of Africa and the Americas. (It is not just Latin America that has figured infrequently in histori- cal sociology; even North America is eclipsed by the usual focus on Europe.) In his monumental book Waves of War (2012), Wimmer used this approach to study war as a sociological phenomenon, drawing on large original data sets. He found that if we look at all wars that occurred throughout history, a major shift has occurred. Prior to the nineteenth century, most wars were driven by conquest, or the desire of states to achieve or throw off a certain balance of power in their region. More recently, wars have been driven by ethnic and nationalist concerns. According to Wimmer, whereas Karl Marx once proclaimed that the twentieth century would be the age of revolution- ary class struggle, it turned into the age of ethno-nationalist conflict. Using a global data set of his own creation, Wimmer found that the existence of nationalist organi- zations in a territory more than doubles the probability of war at any time. It is CONCEPT CHECKS 3 unlikely that a researcher using a quali- tative approach with small numbers of cases would have come up with such a 1. What are the main advantages and limitations of ethnography as a research method? finding, because such an approach would 2. Contrast the two types of questions com- have focused on positive cases (i.e., on a monly used in surveys. handful of nationalist wars), rather than 3. What is a random sample? systematically compared a large number 4. Discuss the main strengths of experiments. of pre-nationalist contexts with nation- alist ones. 4 UNANSWERED QUESTIONS Although the research methods of sociology have become quite advanced over the past 50 years, many open or unanswered questions still cause much disagreement among sociologists. Can Sociology Identify Causes and Effects? One of the main problems faced in research methodology is the analysis of cause and effect. One difficult causal relationship to understand is an association in which one social context produces a certain effect. For example, if you live in a poor neighbor- hood, some sociologists think you are more likely to be unemployed or obese. But is the neighborhood the cause, or is it the other way around? Isn’t it also possible that the kinds of people who would be unemployed or obese tend to gravitate toward living in cer- tain kinds of neighborhoods? Although one of the main tasks of sociology is to identify 48 CHAPTER 2 Asking and Answering Sociological Questions causes and effects, there is a crisis of confi- dence among many scholars that such a goal is empirical investigation Ċ Factual inquiries carried out in any area of sociological study. not as easily attained as once thought. How Can Social Research Avoid Exploitation? All research involving human beings can pose ethical dilemmas. A key question for sociologists is whether research poses risks to subjects that are greater than the risks those subjects face in their everyday lives. For example, ethnographers and field research- ers conducting research in areas with high crime rates potentially risk getting their sub- jects arrested with their writings or getting themselves arrested simply for observing and participating in the lives of the people whom they are trying to understand. Although a great deal of sensitivity to such issues exists in contemporary social science, one largely unan- swered question relates to exploitation. Are social scientists benefitting at their subjects’ expense? Most studies don’t address such questions, and some social scientists are never forced to think seriously about them. The question of exploitation arises more in qualita- tive field studies than in quantitative research, but it must be considered whenever people’s careers come to depend on the advancement of sociological knowledge. For example, if a scholar earns money on a book based on cooperation with research subjects, should that money be shared with the subjects? Can We Really Study Human Social Life in a Scientific Way? To answer this question, we must first understand what science means. Science is the use of systematic methods of empirical investigation, the analysis of data, theoretical thinking, and the logical assessment of arguments to develop a body of knowledge about specific subject matter. According to this definition, sociology is a scientific endeavor. However, sociology is not equivalent to a natural science. Unlike natural phenomena and animals, humans are self-aware beings who confer sense and purpose on what they do. We can’t describe social life accurately unless we grasp the concepts that people apply in their own behavior. For instance, to describe a death as a suicide means knowing what the person intended when he died. Suicide can occur only when an individual has self- destruction actively in mind. If he accidentally steps in front of a car and is killed, he cannot be said to have committed suicide. The fact that we cannot study human beings in the same way as we can study objects in nature is an advantage to sociological researchers, who profit from being able to pose questions directly to those they study: other human beings. In other respects, sociology encounters difficulties not pres- ent in the natural sciences. People who are aware that their activities are being CONCEPT CHECKS 3 scrutinized may not behave normally; 1. How are the ethical dilemmas that social scientists face different from those that other they may consciously or unconsciously researchers encounter in the physical or the portray themselves in a way that differs biological sciences? from their usual attitudes. They may even 2. Why should sociologists be concerned about try to “assist” the researcher by giving the the exploitation of the people they study? responses they believe she wants. Unanswered Questions49 THE BIG PICTURE Chapter 2 Asking and Answering Sociological Questions 1 2 Asking and Answering Basic Sociological Questions: Concepts Historical Context p. 33 p. 37 LEARNING OBJECTIVES Learn the steps of the research Contrast Park’s and Ogburn’s visions process. Name the different types of sociology as a science. of questions sociologists address Understand their influence on in their research—factual, theoretical, contemporary sociological research. comparative, and developmental. TERMS TO KNOW quantitative method` ‘ qualitative method` ‘ Ufpotheses 1. Compare and contrast CONCEPT qualitative and quantitative CHECKS 1. How did Park and Ogburn research methods. approach sociological research 2. What are the seven steps of differently? In what ways the research process? did each sociologist influence 3. What is a hypothesis? researchers today? Exercises: Thinking Sociologically 1. Suppose the dropout rate in your local high school 2. Explain the advantages and disadvantages of increased dramatically. Faced with such a serious documentary research. What will it yield that will problem, the school board offers you a $500,000 grant be better than experimentation, surveys, and to study the sudden increase. Following the ethnographic fieldwork? What are its limitations recommended procedures outlined in the text, explain compared with those approaches? how you would conduct your research. What hypotheses might you test? How would you prove or disprove them? 3 Asking and Answering 4 Sociological Questions Unanswered Today: Research Questions Methods p. 38 p. 48 Familiarize yourself with the methods Understand how research methods available to sociological researchers generate controversies and ethical and recognize the advantages and dilemmas for sociologists. disadvantages of each. research method` ‘ RahnograpUf ‘ participant observNaV\[ ‘ `brvRf ‘ population ‘ pilot studf ‘ sample ‘ representative sample ‘ sampling ‘ random sampling ‘ measures of central tendency ‘ P\rrelation coefficients ‘ mean ‘ mode ‘ median‘ empirical investigation standard decVNaV\[ ‘ QRT_Re of dispersaY ‘ Reperimena ‘ P\mparative research 1. What are the main advantages 1. How are the ethical dilemmas that and limitations of ethnography social scientists face different as a research method? from those that other researchers 2. Contrast the two types of questions encounter in the physical or the commonly used in surveys. biological sciences? 3. What is a random sample? 2. Why should sociologists be 4. Discuss the main strengths of concerned about the exploitation experiments. of the people they study?