-Earl_Babbie-_The_Practice_of_Social_Research(BookFi)

The Politics of Social Research 75 harming subjects clearly relate to Western ideas commitment to making SOciety more humane has about the protection of civil liberties-no one has been called irresponSible. developed a set of political norms that all social researchers accept. In Chapter 10, we'll examine partidpatOlY action research which is explicitly committed to using so- The only partial exception to the lack of politi- cial research for purposes designed and valued by cal norms is the generally accepted view that a re- the subjects of the research. Thus, for example, re- searcher's personal political orientation should not searchers committed to improving the working interfere with or unduly influence his or her sci- conditions for workers at a factory would ask the entific research. It would be considered improper workers to define the outcomes they would like to for a researcher to use shoddy techniques or to dis- see and to have a hand in conducting social re- tort or lie about his or her research as a way of fur- search relevant to achieving the desired ends. The thering the researcher's political views. As you can role of the researchers is to ensure that the workers imagine, however, studies are often enough at- have access to professional research methods. tacked for allegedly violating this norm. Quite aside from abstract disagreements about Objectivity and Ideology whether social science can or should be value-free, many have argued about whether particular re- In Chapter L I suggested that social research can search undertakings are value-free or whether they never be totally objective, because researchers are represent an intrusion of the researcher's own po- human and therefore necessarily subjective. As a litical values. Typically, researchers have denied collective enterprise, science achieves the equiva- such intrusion, and their denials have then been lent of objectivity through intersubjectivity. That is, challenged. Let's look at some examples of the con- different scientists, having different subjective troversies this issue has produced. views, can and should arrive at the same results when they employ accepted research techniques. Social Research and Race Essentially, this will happen to the extent that each can set personal values and views aside for the du- Nowhere have social research and politics been ration of the research. more controversially intertwined than in the area of racial relations. Social researchers studied tlle The classic statement on objectivity and neu- topic for a long time, and the products of the social trality in social science is Max Weber's lecture research have often found their way into practical \"Science as a Vocation\" ([ I 925] 1946). In this talk, politics. A few brief references should illustrate Weber coined the phrase valueJree sociology and the point. urged that sociology. like other sciences, needed to be unencumbered by personal values if it was to In 1896, when the U.S, Supreme Court estab- make a special contribution to society. Liberals and lished the principle of \"separate but equal\" as a conservatives alike could recognize the \"facts\" of means of reconciling the Fourteenth Amendment's social science, regardless of how those facts ac- guarantee of equality to African Americans with corded with their personal politics. the norms of segregation, it neither asked for nor cited social research. Nonetheless, it is widely be- Most social researchers have agreed with this lieved that the Court was influenced by the writ- abstract ideaL but not alL Marxist and neo-Marxist ings of William Graham Surrmer, a leading social scholars, for example, have argued that social sci- scientist of his era. Sumner was noted for his view ence and social action cannot and should not be that the mores and folkways of a society were rela- separated. Explanations of the status quo in society, tively inlpervious to legislation and social planning. they contend, shade subtly into defenses of that His view has often been paraphrased as \"stateways same status quo. Simple explanations of the social do not make folkways.\" Thus, the Court ruled that functions of. say, discrimination can easily become it could not accept the assumption that \"social justifications for its continuance. By the same to- prejudices may be overcome by legislation\" and de- ken, merely studying society and its ills without a nied the wisdom of \"laws which conflict with the

76 ., Chapter 3: The Ethics and Politics of Social Research general sentiment of the community\" (Blaunstein libraries, laboratory facilities, and high expendi- and Zangrando 1970: 308). As many a politician tures per student made little difference. Instead, has said, \"You can't legislate morality.\" Coleman reported that family and neighborhood factors had the most influence on academic When the doctrine of \"separate but equal\" was amievement. overturned in 1954 (BraH'I1 v. Board ofEducation) , the new Supreme Court decision was based in part Coleman's findings were not well received by on the conclusion that segregation had a detrimen- many of the social researchers who had been active tal effect on African American children. In drawing in the civil rights movement. Some scholars criti- that conclusion, the Court cited several sociological cized Coleman's work on methodological grounds, and psychological research reports (Blaunstein and but many others objected hotly on the grounds Zangrando 1970). that the findings would have segregationist political consequences. The controversy that raged around For the most part, social researchers in this cen- the Coleman report was reminiscent of that pro- tury have supported the cause of African American voked a year earlier by Daniel Moynihan (1965) in equality in the United States, and their convictions his critical analysis of the African American family often have been the impetus for their research. in the United States. Moreover, they've hoped that their research will lead to social change. There is no doubt, for exam- Another example of political controversy sur- ple, that Gunnar Myrdal's classic two-volume study rounding social research in connection with race (1944) of race relations in the United States had a concerns IQ scores. In 1969, Arthur Jensen, a significant impact on the topic of his research. Harvard psychologist, was asked to prepare an ar- Myrdal amassed a great deal of data to show that the ticle for the Harvard Educational Review examining position of African Americans directly contradicted the data on racial differences in IQ test results US. values of social and political equality. Further, (Jensen 1969). In the article, Jensen concluded that Myrdal did not attempt to hide his ovvn point of genetic differences between African Americans and view in the matter. (You can pursue Myrdal's land- whites accounted for the lower average IQ scores of mark research further online by searching for African Americans. Jensen became so identified \"Gunnar Myrdal\" or \"An American Dilemma.\") with that position that he appeared on college cam- puses across the country discussing it. Many social researchers have become directly involved in the civil rights movement, some more Jensen's research has been attacked on nu- radically than others. Given the broad support for merous methodological bases. Critics charged that ideals of equality, research conclusions supporting much of the data on which Jensen's conclusion the cause of equality draw little or no criticism. To was based were inadequate and sloppy-there are recognize how solid the general social science many IQ tests, some worse than others. Similarly, position is in this matter, we need only examine a it was argued that Jensen had not taken social- few research projects that have produced conclu- environmental factors sufficiently into account. sions disagreeing with the predominant ideological Other social researchers raised still other method- position. ological objections. Most social researchers have-overtly, at Beyond the scientific critique, however, many least-supported the end of school segregation. condemned Jensen as a racist. Hostile crowds Thus, an immediate and heated controversy arose booed him, drowning out his public presentations. in 1966 when James Coleman, a respected sociolo- Ironically, Jensen's reception by several university gist, published the results of a major national study audiences did not differ significantly from the of race and education. Contrary to general agree- reception received by abolitionists a century before, ment, Coleman found little difference in academic when the prevailing opinion favored leaving the in- performance between African American students stitution of slavery intact. attending integrated scl1001s and those attending segregated ones. Indeed, such obvious things as Many social researchers limited their objections to the Moynihan, Coleman, and Jensen research to

The Politics of Social Research 77 scientific, methodological grounds. The political tend to remain touchy about research on sex. In firestorms ignited by these studies, however, point 1987, the National Institutes of Health (NIH), out how ideology often shows up in matters of so- charged with finding ways to combat the AIDS epi- cial research. Although the abstract model of sci- demic, found they needed hard data on contempo- ence is divorced from ideology, the practice of sci- rary sexual practices if they were to design effective ence is not. anti-AIDS programs. Their request for research proposals resulted in a sophisticated study design To examine a more recent version of the by Edward O. Laumann and colleagues. The pro- controversy surrounding race and achievement, posed study focused on the different patterns of search the web for differing points of view con- sexual activity characterizing different periods of cerning \"The Bell Curve\"-sparked by a book life, and it received rave reviews from the NIH and with that title by Richard 1. Herrnstein and their consultants. Charles Murray. Enter Senator Jesse Helms (R-North Carolina) nJe Politics ofSexual Research and Congressman William Dannemeyer (R- California). In 1989, having learned of the As I indicated earlier, the Laud Humphreys' study Laumann study, Helms and Dannemeyer began a of tearoom trade raised ethical issues that researm- campaign to block the study and shift the same ers still discuss and debate. At the same time, it amount of money to a teen celibacy program. seems clear that much of the furor raised by the Anne Fausto-Sterling, a biologist, sought to under- research was related to the subject matter itself. stand the opposition to the Laumann study. As I have written elsewhere, The surveys, Helms argued, are not really in- Laud Humphreys didn't just study S-E-X but tended \"to stop the spread of AIDS. The real observed and discussed homosexuality. And it purpose is to compile supposedly scientific facts wasn't even the caring-and-cornrnitted- to support the left-wing liberal argument that relationships-between-two-people-who- homosexuality is a normal, acceptable life- just-happen-to-be-of-the-same-sex homosexu- style.... As long as I am able to stand on the ality but tawdry encounters between strangers floor of the US. Senate,\" he added, \"I am never in public toilets. Only adding the sacrifice going to yield to that sort of thing, because it is of Christian babies could have made this not just another life-style; it is sodomy.\" more inflammatory for the great majority of Americans in 1970. (Fallsto-Sterling 1992) (Babbie 2004: 12) Helms won a 66-34 vote in favor of his amend- ment in the US. Senate. Although the House of Whereas Humphreys' researm topic proved Representatives rejected the amendment, and it unusually provocative for many, much tamer sexu- was dropped in conference committee, government ality research has also engendered outcries of pub- funding for the study was put on hold. Laumann lic horror. During the 1940s and 1950s, the biolo- and his colleagues then turned to the private sector gist Alfred Kinsey and his colleagues published and obtained funding, albeit for a smaller study, landmark studies of sexual practices of American from private foundations. Their research results men (1948) and women (1953). Kinsey's extensive were published in 1994 as The Social Organization of interviewing allowed him to report On frequency of Sexuality. sexual activity, premarital and extramarital sex, homosexual behavior, and so forth. His studies Politics and the Census produced public outrage and efforts to close his re- search institute at Indiana University. There is probably a political dimension to every at- tempt to study human social behavior. Consider Although today most people no longer get the matter of the U S. decennial census, mandated worked up about the Kinsey reports, Americans

78 Chapter 3: The Ethics and Politics of Sodal Research by the Constitution. The original purpose was to January 1 was then census day, transient agri- discover the population sizes of the various states to cultural workers were \"incorrectly\" counted in determine their proper representation in the House cities rather than on the farms to which they of Representatives. Whereas each state gets two would return in time for spring planting. (Cen- Senators, large states get more Representatives than sus day was later shifted to April 1, where it has small ones do. So what could be simpler? Just remained.) The arguments dragged out for a de- count the number of people in each state. cade, and Congress was not reapportioned until after the next census. From the beginning, there was nothing simple about counting heads in a dispersed, na- (Prewitt 2003) tional population like the United States. Even the definition of a \"person\" was anything but straight- In more recent years, concern for undercount- forward. A slave, for example, counted as only ing the urban poor has become a political issue. The three-fourths of a person for purposes of the big cities, which have the most to lose from the census. This decreased the representation of the undercounting, typically vote Democratic rather slave-holding Southern states, though counting than Republican, so you can probably guess which slaves as whole people might have raised the dan- party supports efforts to improve the counting and gerously radical idea that they should be allowed which party is less enthusiastic. By the same token, to vote. when social scientists have argued in favor of re- placing the attempt at a total enumeration of the Further, the logistical problems of counting population vvith modern survey sampling methods people who reside in suburban tract houses, urban (see Chapter 7 for more on sampling), they have apartments, college dorms, military barracks, enjoyed more support from Democrats, who would farms, cabins in the woods, and illegal housing stand to gain from such a methodological shift, units, as well as counting those who have no place than from Republicans, who would stand to lose. to live, has always presented a daunting task It's Rather than suggesting Democrats support science the sort of challenge social researchers tackle with more than Republicans do, this situation offers relish. However, the difficulty of finding the hard- another example of how the political context in to-reach and the techniques created for doing so which we live and conduct social research often cannot escape the political neL affects that research. Kenneth Prewitt, who directed the Census Bu- Politics with aLittle lip\" reau from 1998 to 200 L describes some of the po- litical aspects of counting heads: Social research is often confounded by political ide- ologies, but the \"politics\" of social research runs far Between 1910 and 1920, there was a massive deeper stilL Social research in relation to contested wartime population movement from the rural, social issues simply cannot remain antiseptically Southern states to industrial Northern cities. In objective-particularly when differing ideologies 1920, for the first time in American history, the are pitted against each other in a field of social census included more city dwellers than rural science data. residents. An urban America was something new and disturbing, especially to those who The same is true when research is invoked in held to the Jeffersonian belief that independent disputes between people with conflicting interests. farmers best protected democracy. Among those For instance, social researchers who have served as of this persuasion were rural, conservative con- \"expert witnesses\" in court would probably agree gressmen in the South and West. They saw that that the scientific ideal of a \"search for truth\" seems reapportionment would shift power to factory- hopelessly naive in a trial or lawsuit. Although ex- based unions and politically radical immigrants pert witnesses technically do not represent either concentrated in Northeastern cities. Conserva- side in court, they are, nonetheless, engaged by tives in Congress blocked reapportionment, only one side to appear, and their testimony tends complaining among other things that because

The Politics of SodaI Research 79 to support the side of the party who pays for their social researchers are going to act like human be- time. This doesn't necessarily mean that these wit- ings, and we must take this into account in assess- nesses will lie on behalf of their patrons, but the ing their findings. This recognition does not in- contenders in a lawsuit are understandably more validate their research or provide an excuse for likely to pay for expert testimony that supports rejecting findings we happen to dislike, but it does their case than for testimony that attacks it. need to be taken into account. Thus, as an expert witness, you appear in court Politics in Perspective only because your presumably scientific and honest judgment happens to coincide with the interests of Although the ethical and the political dimensions the party paying you to testify. Once you arrive in of research are in principle distinct they do inter- court and swear to tell the truth, the whole truth, sect. Whenever politicians or the public feel that and nothing but the truth, however, you find your- social research is violating ethical or moral stan- self in a world foreign to the ideals of objective con- dards, they'll be quick to respond with remedies of templation. Suddenly, the norms are those of win- their own. Moreover, the standards they defend ning and losing. As an expert witness, of course, all may not be those of the research community. Even you have to lose is your respectability (and perhaps when researchers support the goals of measures the chance to earn fees as an expert witness in the directed at the way research is done, the means future). Still, such stakes are high enough to create specified by regulations or legislation can hamstring discomfort for most social researchers. research. I recall one case in federal court when I was tes- Legislators show special concern for research tifying on behalf of some civil service workers who on children. Although the social research norms had had their cost-of-living allowance (COLA) cut discussed in this chapter would guard against on the basis of research I thought was rather shoddy. bringing any physical or emotional harm to chil- I was engaged to conduct more \"scientific\" research dren, some of the restrictive legislation introduced that would demonstrate the injustice worked from time to time borders on the actions of one against the civil servants (Babbie 1982: 232-43). particular western city, which shall remain name- less. In response to concerns that a public school I took the stand, feeling pretty much like a re- teacher had been playing New Age music in class spected professor and textbook author. In short and encouraging students to meditate, the city order, however, I found I had moved from the council passed legislation stating that no teacher academy to the hockey rink. Tests of statistical could do anything that would \"affect the minds of significance and sampling error were suddenly less students\"! relevant than a slap shot. At one point, an attorney from Washington lured me into casually agreeing I hope you take away four main lessons from that I was familiar with a certain professional jour- this discussion. First, science is not untouched by nal. Unfortunately, the journal did not exist. I was politics. The intrusion of politics and related ideolo- mortified and suddenly found myself shifting do- gies is not unique to social research; the natural sci- mains. Without really thinking about it, I now was ences have experienced and continue to experience less committed to being a friendly Mr. Chips and similar situations. But social science, in particular, is more aligned with ninja-professor. I would not be a part of social life. Social researchers study things fully satisfied until 1, in turn, could mortify the at- that matter to people-things they have firm, per- torney, which I succeeded in doing. sonal feelings about and things that affect their lives. Moreover, researchers are human beings, and Even though the civil servants got their cost-of- their feelings often show through in their profes- living allowance back, I have to admit I was also sionallives. To think otherwise would be naive. concerned with how I looked in front of the court- room assemblage. I tell you this anecdote to illus- Second, science does proceed in the midst of trate the personal \"politics\" of human interactions political controversy and hostility. Even when involving presumably scientific and objective researchers get angry and call each other names, or research. We need to realize that as human beings

80 \" Chapter 3: The Ethics and Politics of Sodal Research when the research community comes under attack Ethical Issues in Social Research from the outside, scientific inquiry persists\" Studies are done, reports are published, and new things & What is ethical and unethical in research is ulti- are learned\" In short, ideological disputes do not matelya matter of what a community of people bring science to a halt, but they do make it more agree is right and wrong\" challenging-and exciting\" & Researchers agree that participation in research Third, an awareness of ideological considera- should normally be voluntary. This norm, how- tions enriches the study and practice of social re- ever, can conflict with the scientific need for search methods. Many of the established character- generalizability. istics of science, such as intersubjectivity, function to cancel out or hold in check our human short- & Researchers agree that research should not comings, especially those we are unaware of. Oth- harm those who participate in it, unless they erwise, we might look into the world and never see give their informal consent, thereby willingly anything but a reflection of our personal biases and and knowingly accepting the risks of harm. beliefs. & Whereas anonymity refers to the situation in Finally, whereas researchers should not let which even the researcher cannot identify their own values interfere with the quality and specific information with the individuals it de- honesty of their research, this does not mean that scribes, confidentiality refers to the situation in researchers cannot or should not participate in which the researcher promises to keep infor- public debates and express both their scientific mation about subjects private. The most expertise and personal values. You can do scientifi- straightforward way to ensure confidentiality is cally excellent research on racial prejudice, all the to destroy identifying information as soon as it's while being opposed to prejudice and saying so. no longer needed. Some would argue that social scientists, because of their scientific expertise in the workings of society, & Many research designs involve a greater or have an obligation to speak out, rather than leaving lesser degree of deception of subjects. Because that role to politicians, journalists, and talk-show deceiving people violates common standards of hosts. Herbert Gans (2002) writes of the need for ethical behavior, deception in research requires \"public sociologists\": a strong justification-and even then the justification may be challenged. A public sociologist is a public intellectual who applies sociological ideas and findings to social & Social researchers have ethical obligations to the (defined broadly) issues about which sociology community of researchers as well as to subjects. (also defined broaclly) has something to say These obligations include reporting results fully Public intellectuals comment on whatever is- and accurately as well as disclosing errors, limi- sues show up on the public agenda; public soci- tations, and other shortcomings in the research. ologists do so only on issues to which they can apply their sociological insights and findings. & Professional associations in several disciplines publish codes of ethics to guide researchers. MAIN POINTS These codes are necessary and helpful, but they do not resolve all ethical questions. Introduction & In addition to tedmical, scientific considera- Two Ethical Controversies tions, social research projects are likely to be & Laud Humphreys' study of \"tearoom\" encoun- shaped by administrative, ethical, and political ters and Stanley Milgram's study of obedience considerations. raise ethical issues that are debated to this day\" The Politics of Social Research & Social research inevitably has a political and ideological dimension. Although science is neutral on political matters, scientists are not.

Moreover, much social research inevitably in- Review Questions and Exercises 81 volves the political beliefs of people outside the research community. risk arrest as an accomplice after the fact, the researcher complies. & Although most researchers agree that political orientation should not unduly influence re- c After completing the final draft of a book search, in practice separating politics and ideol- reporting a research project, the researcher- ogy from the conduct of research can be quite author discovers that 25 of the 2,000 survey difficult. Some researchers maintain that re- interviews were falsified by interviewers\" To search can and should be an instrument of so- protect the bulk of the research, the author cial action and change. More subtly, a shared leaves out this information and publishes ideology can affect the way other researchers the book receive one's research. d. Researchers obtain a list of right-wing radi- & Even though the norms of science cannot force cals they wish to study They contact the individual researchers to give up their personal radicals with the explanation that each has values, the intersubjective character of science been selected \"at random\" from among the provides a guard against scientific findings be- general population to take a sampling of ing the product of bias only\" \"public opinion.\" KEY TERMS e\" A college instructor who wants to test the effect of unfair berating administers an hour The following terms are defined in context in the exam to both sections of a specific course. chapter and at the bottom of the page where the The overall performance of the two sections term is introduced, as well as in the comprehensive is essentially the same. The grades of one glossary at the back of the book section are artificially lowered, however, and the instructor berates the students for anonymity debriefing performing so badly\" The instructor then ad- confidentiality informed consent ministers the same final exam to both sec- tions and discovers that the performance of REVIEW QUESTIONS AND EXERCISES the unfairly berated section is worse. The hypothesis is confirmed, and the research I. Consider the following real and hypothetical re- report is published. search situations. What is the ethical component in each example? How do you feel about it? Do £. In a study of sexual behavior, the investiga- you think the procedures described are ulti- tor wants to overcome subjects' reluctance mately acceptable or unacceptable? You might to report what they might regard as shame- find it useful to discuss some of these situations ful behavioL To get past their reluctance, with classmates. subjects are asked, \"Everyone masturbates now and then; about how much do you a. A psychology instructor asks students in an masturbate?\" introductory psychology class to complete questionnaires that the instructor will ana- g\" A researcher studying dorm life on campus lyze and use in preparing a journal article discovers that 60 percent of the residents for publication. regularly violate restrictions on alcohol con- sumption. Publication of this finding would b\" After a field study of deviant behavior dur- probably create a furor in the campus com- ing a riot, law enforcement officials demand munity. Because no extensive analysis of al- that the researcher identify those people cohol use is planned, the researcher decides who were observed looting. Rather than to keep this finding quiet. h. To test the extent to which people may try to save face by expressing attitudes on mat- ters they are wholly uninformed about, the researcher asks for their attitudes regarding a fictitious issue. L A research questionnaire is circulated among students as part of their university registra- tion packet. Although students are not told they must complete the questionnaire, the

82 Chapter 3: The Ethics and Politics of Social Research hope is that they will believe they must- lee, Raymond. 1993. Doillg Research 011 Sensitive thus ensuring a higher completion rateo Topics. Newbury Park, CA: Sage. This book examines the conflicts between scientific re- j. A researcher pretends to join a radical search needs and the rights of the people political group in order to study it and is involved-with guidelines for dealing with such successfully accepted as a member of the conflicts. inner planning circle. What should the re~ searcher do if the group makes plans for the Sweet, Stephen. 19990 \"Using a Mock Institutional following? Review Board to Teach Ethics in Sociological Research.\" Teachillg Sociology 27 (January): (1) A peaceful, though illegal, demonstration 55-59. Though written for professors, this ar- ticle provides some research examples that chal- (2) The bombing of a public building during lenge your ethical instincts. a time it is sure to be unoccupied SPSS EXERCISES (3) The assassination of a public official See the booklet that accompanies your text for exer- 2. Review the discussion of the Milgram experi- cises using SPSS (Statistical Package for the Social Sci- ment on obedienceo How would you design a ences). There are exercises offered for each chapter, study to accomplish the same purpose while and you'll also find a detailed primer on using SPSS. avoiding the ethical criticisms leveled at Mil- gram? Would your design be equally valid? Online Study Resources Would it have the same effect? Sociology.~ Now'\": Research Methods 3. Suppose a researcher who is personally in favor of small families-as a response to the problem 10 Before you do your final review of the chapter, of overpopulation-wants to conduct a survey take the SociologyNow~ Research lvlethods diagnos- to determine why some people want many tic quiz to help identify the areas on which you children and others don't. What personal- should concentrate. You'll find information on involvement problems would the researcher this online tool, as well as instructions on how face, and how could she or he avoid them? to access all of its great resources, in the front of What ethical issues should the researcher take the book. into account in designing the survey? 20 As you review, take advantage of the Sociology 4. Using InfoTrac College Edition, search for \"in- Now: Research Methods customized study plan, formed content\" and then narrow your search based on your quiz results . Use this study plan to \"researcho\" Skim the resulting articles and with its interactive exercises and other re- begin to identify groups of people for whom in- sources to master the materiaL formed consent may be problematic-people who may not be able to give it. Suggest some 3. When you're finished with your review, take ways in which the problem might be overcome. the posttest to confirm that you're ready to move on to the next chapteL ADDITIONAL READINGS WEBSITE FOR THE PRACTICE Hamnett, Michael E, Douglas J. Porter, Amarjit OF SOCIAL RESEARCH 11 TH EDITION Singh, and Krishna Kumar- 1984. Ethics, Politics, and International Social Science ResearcJL Hon- Go to your book's website at http://sociology olulu: University of Hawaii Press. Discussions of .wadsworthcom/babbie_practice lIe for tools to research ethics typically focus on the interests of aid you in studying for your exams. You'll find Tutorial the individual participants in research projects, Quizzes with feedback, Internet Exercises, Flashcards, and but this book raises the level of the discussion to include the rights of whole societies. Homan, RogeL 1991. Tlze Ethics of Social Research london: Longman. A thoughtful analysis of the ethical issues of social science research, by a practicing British social researcher,

Online Study Resources 83 Chapter Tutorials, as well as E\\1ended Projeas, InfoTrae Department of Health and Human Services, College Edition search terms, Serial Research in C)berspace, Protection of Human Subjects (45 CFR Part 46) GSS Data, y~;;b Links, and primers for using various hnp:llwww.hhs . gov/ohrp/humansubjects/ data-analysis software such as SPSS and NVivo. guidance/45cfr46.htm Here is the primary federal regulation governing the WEB LINKS FOR THIS CHAPTER treatment of human subjects and providing a basis for the actions of institutional review boards. Please realize that the Internet is an evolv- ing entity, subject to change. Nevertheless, Research Council of NOIway, Guidelines for these few websites should be fairly stable. Research Ethics in the Social Sciences, Law Also, check your book's website for even more H0b and the Humanities Links http://www.etikkom.no/Engelsk/J\\TESH This report, by the National Conunittee for Research Anlerican Sociological Association, Code Ethics in the Social Sciences and the Humanities, pro- of Ethics vides an in-depth examination of the topic http://www.asanetorg/page. vvw?section=Ethics& name=Ethics Most professional associations have codes of ethics intended to guide the activities of their members. This one is a good illustration of the genre.

osing problems properly is often more difficult than answering them. Indeed, aproperly phrased question often seems to answer itself. .•You may have discovered the answer to aquestion just In the process of making the question clear to someone else. Part 2deals with what should be observed; that is, Part 2considers the posing of proper scientific questions, the structuring of inquiry. Part 3will describe some ofthe specific methods of social scientific observation. Chapter 4 addresses the beginnings of research. It examines some of the purposes of inquiry, units of analysis, and the reasons scientists get involved in re- search projects. Chapter 5'deals with the specification of what it is you Want to measure-the processes of conceptualiza- tion and operationalization. It looks at some of the terms that you and I use quite casually in everyday life-prejudic~ liberalism, happjnes~ and so forth-and

I I shows how essential it is to clarify what we really mean composite measure of gender-based egalitarianism. by such terms when we do research. 'This process of Although such composite measures are constructed dur- clarification is called conceptualization. ing the analysis of data (see Part 4), the raw materials for them must be provided for in the design and execu- Once we clarify what we mean by certain terms, we tion of data colle(tion. can then measure the referents of those terms. The pro- cess of devising steps or operations for measuring what Finally, we'll look at how social researchers sele(t we want to study is called operationalization. Chapter 5 people or things for observation. Chapter 7, on sam- deals with the topic of operationalization in general, pling, addresses the fundamental scientific issue of gen- paying special attention to the framing of questions for eralizability. As you'll see, we can select afew people or interviews and questionnaires. things for observation and then apply what we observe to amuch larger group. For example, by surveying To complete the introduction to measurement, 2,000 U.s. citizens about whom they favor for president Chapter 6 breaks with the chronological discussion of of the United States, we can accurately predkt how tens how research is conducted. In this chapter, we'll exam- of millions will vote. In this chapter, we'll examine tech- ine techniques for measuring variables in quantitative niques that increase the generalizability of what we research through the combination of several indicators: observe. indexes, scales, and typologies. As an example, we might ask survey respondents five different questions What you learn in Part 2 will bring you to the verge about their attitudes toward gender equality and of making controlled social scientific observations. Part then combine the answers to all five questions into a 3 will then show you how to take that next step. 85

Research Design Introduction The Time Dimension Three Purposes of Research Cross-Sectional Studies Longitudinal Studies Exploration Approximating Longitudinal Description Studies Explanation Examples of Research Strategies The Logic of Nomothetic Explanation How to Design a Research Project Criteria for Nomothetic Causality Getting Started False Criteria for Nomothetic Conceptualization Causality Choice of Research Method Operationalization Necessary and Sufficient Causes Population and Sampling Observations Units of Analysis Data Processing Individuals Analysis Groups Application Organizations Research Design in Review Social Interactions Social Artifacts The Research Proposal Units of Analysis in Review Faulty Reasoning about Units Elements of a Research of Analysis: The Ecological Proposal Fallacy and Reductionism Sociology~ NOW'\": Research Methods Use this online tool to help you make the grade on your next exam. After reading this chapter, go to the \"Online Study Resources\" at the end of the chapter for instmctions on how to benefit from SodologyNow: Research lyre/hods.

Three Purposes of Research 87 Introduction of terrorists, case studies of terrorist organizations, case studies on types of terrorism, case studies on Science is an enterprise dedicated to \"finding out.\" particular terrorist incidents, and case studies of No matter what you want to find out, though, there terrorism in selected regions and countries.\" (2004: will likely be a great many ways of doing it. That's 27) Quantitative researchers, on the other hand, true in life generally. Suppose, for example, that you addressed terrorism in a variety of ways, including want to find out whether a particular automobile- analyses of media coverage, statistical modeling of say, the new Burpo-Blasto-would be a good car terrorist events, and the use of various databases for you. You could, of course, buy one and find out relevant to the topic. As you'll see in this chapter, that way. Or you could talk to a lot of B-B owners or any research topic can be approached from many to people who considered buying one but didn't. different directions. You might check the classified ads to see if there are a lot of B-Bs being sold cheap. You could read a con- This chapter provides a general introduction to sumer magazine evaluation of Burpo-Blastos. A research design, whereas the other chapters in Part similar situation occurs in scientific inquiry. 2 elaborate on specific aspects of it. In practice, all aspects of research design are interrelated. As you Ultimately, scientific inquiry comes down to read through Part 2, the interrelationships among making observations and interpreting what you've parts will become clearer. observed, the subjects of Parts 3 and 4 of thls book. Before you can observe and analyze, however, you We'll start by briefly examining the main pur- need a plan. You need to determine what you're poses of social research. Then, we'll consider units going to observe and analyze: why and how. That's of analysis-the what or whom you want to study. what research design is all about. Next we'll consider ways of handling time in social research, or how to study a moving target that Although the details vary according to what changes over time. you wish to study, you face two major tasks in any research design. First you must specify as clearly as With these ideas in hand, we'll turn to how to possible what you want to find out. Second, you design a research project. This overview of the re- must determine the best way to do it. Interestingly, search process serves two purposes: Besides describ- if you can handle the first consideration fully, you'll ing how you might go i3bout designing a study, it probably handle the second in the same process. As provides a map of the remainder of this book. mathematicians say, a properly framed question contains the answer. Finally, we'll look at the elements of research proposals. Often, the actual conduct of research Let's say you're interested in conducting social needs to be preceded by a detailing of your inten- research on terrorism. When Jeffrey Ross (2004) tions-to obtain funding for a major project or per- addressed this issue, he found the existing studies haps to get your instructor's approval for a class used a variety of qualitative and quantitative ap- project. You'll see that the research proposal pro- proaches. Qualitative researchers, for example, vides an excellent opportunity for you to consider generated original data through all aspects of your research in advance. Autobiographies Three Purposes of Research Incident Reports and Accounts Social research can serve many purposes. Three of the most common and useful purposes are explo- Hostages' EX'Periences with Terrorists ration, description, and explanation. Although a given study can have more than one of these Firsthand Accounts of Implementing Policies purposes-and most do-examining them Ross goes on to discuss some of the secondary ma- terials used by qualitative researchers: \"biographies

88 Chapter 4: Research Design separately is useful because each has different A while back, for example, I became aware of implications for other aspects of research design. the growing popularity of something called \"chan- neling,\" in which a person known as a channel or Exploration medium enters a trance state and begins speaking with a voice that claims it originates outside the Much of social research is conducted to e;..'{)lore a channel. Some of the voices say they come from a topic, that is, to start to familiarize a researcher with spirit world of the dead, some say they are from that topic. This approach typically occurs when a other planets, and still others say they exist in di- researcher examines a new interest or when the mensions of reality difficult to explain in ordinary subject of study itself is relatively new. human terms. As an example, let's suppose that widespread The channeled voices, often referred to as enti- taxpayer dissatisfaction with the government erupts ties, sometimes use the metaphor of radio or televi- into a taxpayers' revolt. People begin refusing to sion for the phenomenon they represent. \"When pay their taxes, and they organize themselves you watch the news,\" one told me in the course of around that issue. You might like to learn more an interview, \"you don't believe Dan Rather is re- about the movement: How widespread is it? What ally inside the television set. The same is true of levels and degrees of support are there within the me. I use this medium's body the way Dan Rather community? How is the movement organized? uses your television set.\" What kinds of people are active in it? An explo- ratory study could help you find at least approxi- The idea of channeling interested me from sev- mate answers to some of these questions, You eral perspectives, not the least of which was the might check figures vvith tax-collecting officials, methodological question of how to study scientifi- collect and study the literature of the movement, cally something that violates so much of what we attend meetings, and interview leaders. take for granted, including scientific staples such as space, time, causation, and individuality. Exploratory studies are also appropriate for more persistent phenomena. Suppose you're un- Lacking any rigorous theory or precise expecta- happy with your college's graduation requirements tions, I merely set out to learn more. Using some of and want to help change them. You might study the techniques of qualitative field research dis- the history of such requirements at the college and cussed in Chapter 10, I began amassing informa- meet with college officials to learn the reasons for tion and forming categories for making sense of the current standards. You could talk to several stu- what I observed. I read books and articles about the dents to get a rough idea of their sentiments on the phenomenon and talked to people who had at- subject. Though this last activity would not neces- tended channeling sessions. I then attended chan- sarily yield an accurate picture of student opinion, neling sessions myself, observing those who at- it could suggest what the results of a more exten- tended as well as the channel and entity Next, I sive study might be. conducted personal interviews vvith numerous channels and entities. Sometimes exploratory research is pursued through the use of focus groups, or guided small- In most interviews, I began by asking the hu- group discussions. This technique is frequently man channels questions about how they first began used in market research; we'll examine it further in channeling, what it was like, and why they contin- Chapter 10. ued, as well as standard biographical questions. The channel would then go into a trance, whereby the Exploratory studies are most typically done for interview continued vvith the entity speaking. three purposes: (1) to satisfy the researcher's Cli- \"Who are you?\" I might ask. \"Where do you come riosity and desire for better understanding, (2) to from?\" \"Why are you doing this?\" \"How can I tell test the feasibility of undertaking a more extensive if you are real or a fake?\" Although I went into study, and (3) to develop the methods to be em- these interview sessions with several questions pre- ployed in any subsequent study. pared in advance, each of the interviews followed

Three Purposes of Research 89 whatever course seemed appropriate in light of the people you study in your exploratory research may answers given. not be typical of the larger population that interests you. Once you understand representativeness, This example of exploration illustrates where you'll be able to know whether a given exploratory social research often begins. Whereas researchers study actually answered its research problem or working from deductive theories have the key vari- only pointed the way toward an answer. (Repre- ables laid out in advance, one of my first tasks was sentativeness is discussed at length in Chapter 7.) to identify some of the possibly relevant variables. For example, I noted a channel's gender, age, edu- Description cation, religious background, regional origins, and previous participation in things metaphysical. I A major purpose of many social scientific studies is chose most of these variables because they com- to describe situations and events. The researcher monly affect behavior. observes and then describes what was observed. Because scientific observation is careful and delib- I also noted differences in the circumstances of erate, however, scientific descriptions are typically channeling sessions. Some channels said they must more accurate and precise than casual ones are. go into deep trances, some use light trances, and others remain conscious. Most sit down while The US. Census is an excellent example of de- channeling, but others stand and walk about. Some scriptive social research. The goal of the census is to channels operate under pretty ordinary conditions; describe accurately and precisely a wide variety of others seem to require metaphysical props such as characteristics of the U.S. population, as well as the dim lights, incense, and chanting. Many of these populations of smaller areas such as states and differences became apparent to me only in the counties, Other examples of descriptive studies are course of my initial observations. the computation of age-gender profiles of popula- tions done by demographers, the computation of Regarding the entities, I have been interested in crime rates for different cities, and a product- classifying where they say they come from. Over marketing survey that describes the people who the course of my interviews, I've developed a set use, or would use, a particular product. A re- of questions about specific aspects of \"reality,\" at- searcher who carefully chronicles the events that tempting to classify the answers they give. Simi- take place on a labor union picket line has, or at larly, I ask each to speak about future events, least serves, a descriptive purpose. A researcher who computes and reports the number of times Over the course of this research, my examina- individual legislators voted for or against organized tion of specific topics has become increasingly fo- labor also fulfills a descriptive purpose. cused as I've identified variables that seem worth pursuing: gender, education, and religion, for ex- Many qualitative studies aim primarily at de- ample. Note, however, that I began vvith a reason- scription. An anthropological ethnography, for ex- ably blank slate. ample, may try to detail the particular culture of some preliterate society. At the same time, such Exploratory studies are quite valuable in social studies are seldom limited to a merely descriptive scientific research. They're essential whenever a re- purpose. Researchers usually go on to examine searcher is breaking new ground, and they almost why the observed patterns exist and what they always yield new insights into a topic for research. imply. Exploratory studies are also a source of grounded theory, as discussed in Chapter L Explanation The chief shortcoming of exploratory studies The third general purpose of social scientific re- is that they seldom provide satisfactory answers search is to explain things. Descriptive studies to research questions, though they can hint at answer questions of what, where, when, and the answers and can suggest which research meth- ods could provide definitive ones. The reason ex- ploratory studies are seldom definitive in them- selves has to do with representativeness; that is, the

90 Chapter 4: Research Design how; explanatory questions, of why, So when The logic William Sanders (1994) set about describing the of Nomothetic Explanation varieties of gang violence, he also wanted to re- construct the process that brought about violent The preceding examination of what factors might episodes among the gangs of different ethnic cause attitudes about legalizing marijuana illus- groups, trates nomothetic explanation, as discussed in Chapter L Recall that in this model, we try to find Reporting the voting intentions of an electorate a few factors (independent variables) that can ac- is descriptive, but reporting why some people plan count for many of the variations in a given phe- to vote for Candidate A and others for Candidate B nomenon, This explanatory model stands in con- is explanatory, Identifying variables that explain trast to the idiographic model, in which we seek a why some cities have higher crime rates than oth- complete, in-depth understanding of a single case, ers involves explanation, A researcher who sets out to know why an antiabortion demonstration ended In our example, an idiographic approach would in a violent confrontation with police, as opposed suggest all the reasons that one person was op- to simply describing what happened, has an ex- posed to legalization-involving what her parents, planatory purpose, teachers, and clergy told her about it; any bad ex- periences experimenting vvith it; and so forth, Let's look at a specific case, What factors When we understand something idiographically, do you suppose might shape people's attitudes we feel we really understand it. When we know all toward the legalization of marijuana? To an- the reasons why someone opposed legalizing mari- swer this, you might first consider whether men juana, we couldn't imagine that person having any and women differ in their opinions, An explana- other attitude, tory analysis of the 2002 General Social Survey (GSS) data indicates that 38 percent of men and In contrast, a nomothetic approach might sug- 30 percent of women said marijuana should be gest that overall political orientations account for legalized, much of the difference of opinion about legalizing marijuana, Because this model is inherently proba- What about political orientation? The GSS data bilistic, it is more open than the idiographic model show that 55 percent of liberals said marijuana to misunderstanding and misinterpretation, Let's should be legalized, compared with 29 percent of examine what social researchers mean when they moderates and 27 percent of conservatives, Fur- say one variable (nomothetically) causes another. ther, 41 percent of Democrats, compared with Then, we'll look at what they don't mean, 34 percent of Independents and 28 percent of Republicans, supported legalization, Criteria for Nomothetic Causality Given these statistics, you might begin to de- There are three main criteria for nomothetic causal velop an e;.:planation for attitudes toward mari- relationships in social research: (1) the variables juana legalization, Further study of gender and must be correlated, (2) the cause takes place before political orientation might then lead to a deeper the effect, and (3) the variables are nonspurious, explanation of these attitudes, Correlation correlation An empirical relationship betvveen two variables such that (1) changes in one are asso- Unless some actual relationship-or correlation- ciated with changes in the other or (2) particular is found between two variables, we can't say that a attributes of one variable are associated with par- causal relationship exists, Our analysis of GSS data ticular attributes of the other. Correlation in and suggested that political orientation was a cause of of itself does not constitute a causal relationship attitudes about legalizing marijuana, Had the same between the two variables, but it is one criterion percentage of liberals and conservatives supported of causality,

The logic of Nomothetic Explanation 91 legalization, we could hardly say that political ori- and cities: the more mules, the fewer Ph,D:s and entations caused the attitude, Though this cTiterion vice versa, Perhaps you can think of another vari- is obvious, it emphasizes the need to base social re- able that would explain this apparent relationship, search assertions on actual observations rather than The answer is rural versus urban settings. There are assumptions, more mules (and fewer Ph,D:s) in rural areas, whereas the opposite is true in cities, Time Order Or, consider the positive correlation between Next, we can't say a causal relationship exists unless shoe size and math ability among schoolchildren, the cause precedes the effect in time, Notice that it Here, the third variable that explains the puzzling makes more sense to say that most children's reli- relationship is age. Older children have bigger feet gious affiliations are caused by those of their par- and more highly developed math skills, on average, ents than to say that parents' affiliations are caused than younger children do. See Figure 4-1 for an by those of their children-even though it would illustration of this spurious relationship, Observed be possible for you to change your religion and for associations are indicated vvith thin arrows; causal your parents to follow suit, Remember, nomothetic relationships with thick ones. Notice that observed explanation deals vvith \"most cases\" but not all, associations go in both directions, That is, as one variable occurs or changes, so does the other. In our marijuana example, it would make sense to say that gender causes, to some extent, at- The list goes on. Areas with many storks have titudes toward legalization, whereas it would make high birth rates. Those with few storks have low no sense to say that opinions about marijuana de- birth rates, Do storks really deliver babies? Birth termine a person's gender. Notice, however, that rates are higher in the country than in the city; the time order connecting political orientations and more storks live in the country than the city. The attitudes about legalization is less clear, though we third variable here is urban/1'lIral areas. sometimes reason that general orientations cause specific opinions, And sometimes our analyses in- Finally, the more fire trucks that put out a fire, volve two or more independent variables that were the more damage to the structure, Can you guess established at the same time: looking at the effects what the third variable is? In this case, it's the size of gender and race on voting behavior, for ex- ofthe fire, ample, As we'll see in the next chapter, the issue of time order can be a complex matter. Thus, when social researchers say there is a causal relationship between, say, education and Nonspurious racial tolerance, they mean (1) there is a statistical correlation between the two variables, (2) a per- The third requirement for a causal relationship is son's educational level occurred before their cur- that the effect cannot be explained in terms of rent level of tolerance or prejudice, and (3) there is some third variable, For example, there is a correla- no third variable that can explain away the ob- tion between ice-cTeam sales and deaths due to served correlation as spurious. drowning: the more ice cream sold, the more drownings, and vice versa, There is, however, no False Criteria for Nomothetic Causality direct link between ice CTeam and drowning, The third variable at work here is season or temperature, Because notions of cause and effect are well en- Most drowning deaths occur during summer-the trenched in everyday language and logic, it's peak period for ice-CTeam sales, important to specify some of the things social Here are a couple of other examples of spuri- spurious relationship A coincidental statistical ous relationships, or ones that aren't genuine, correlation between two variables, shown to be There is a negative relationship between the num- caused by some third variable, ber of mules and the number of Ph,D:s in towns

92 Chapter 4: Research Design Observed Correlation Positive (direct) correlation ~~...~Shoe size ... Math skill Bigger shoe size is associated with greater math skill, and vice versa. Spurious causal relationships Actual causal relationships Math skill Shoe size Math skill '\" Shoe size Neither shoe size nor math skill is a cause The underlying variable of age causes both of the otheL bigger shoe size and greater math skill, thus explaining the observed correlation. FIGURE 4-1 An Example of aSpurious Causal Relationship. Finding an empirical correlation between two variables does not necessarily estab- lish acausal relationship. Sometimes the observed correlation is the incidental result of other causal relationships, involving other variables. researchers do not mean when they speak of causal consistently found that women are more religious relationships. When they say that one variable than men in the United States. Thus, gender may causes another, they do not necessarily mean to be a cause of religiosity, even if your uncle is a reli- suggest complete causation, to account for excep- gious zealot or you know a woman who is an tional cases, or to claim that the causation exists in avowed atheist. Those exceptional cases do not a majority of cases. disprove the overall, causal pattern. Complete Causation Majority of Cases Whereas an icliographic explanation of causation Causal relationships can be true even if they don't is relatively complete, a nomothetic explanation apply in a majority of cases. For example, we say is probabilistic and usually incomplete. As we've that c11ildren who are not supervised after school seen, social researchers may say that political orien- are more likely to become delinquent than those tations cause attitudes toward legalizing marijuana who are supervised are; hence, lack of supervision even though not all liberals approve nor all conser- is a cause of delinquency. This causal relationship vatives disapprove. Thus, we say that political ori- holds true even if only a small percentage of those entation is one of the causes of the attitude, but not not supervised become delinquent. As long as they the only one. are more likely than those who are supervised to be delinquent, we say there is a causal relationship. Exceptional Cases The social scientific view of causation may vary In nomothetic explanations, exceptions do not from what you are accustomed to, because people disprove a causal relationship. For example, it is commonly use the term cause to mean something

Necessary and Sufficient Causes 93 FIGURE 4-2 Necessary Cause. Being female is anecessary cause of pregnancy; that is, you can't get pregnant unless you are female. that completely causes another thing. The some- cause is the only possible cause of a particular effect. what different standard used by social researchers For example, skipping an exam in this course can be seen more clearly in terms of necessary and would be a sufficient cause for failing it, though sufficient causes. students could fail it other ways as well. Thus, a cause can be sufficient, but not necessary. Necessary Figure 4- 3 illustrates the relationship between and Sufficient Causes taking or not taking the exam and either passing or failing it. A necessmy calise represents a condition that must be present for the effect to follow. For example, it is The discovery of a cause that is both necessary necessary for you to take college courses in order to and sufficient is, of course, the most satisfying out- get a degree. Take away the courses, and the degree come in research. If juvenile delinquency were the never follows. However, simply taking the courses effect under examination, it would be nice to dis- is not a sufficient cause of getting a degree. You cover a single conclition that (I) must be present for need to take the right ones and pass them. Simi- delinquency to develop and (2) always results in larly, being female is a necessary condition of be- delinquency. In such a case, you would surely feel coming pregnant, but it is not a sufficient cause. that you knew precisely what caused juvenile Otherwise, all women would get pregnant. delinquency. Figure 4-2 illustrates this relationship between Unfortunately, we never discover single causes the variables of gender and pregnancy as a matrix that are absolutely necessary and absolutely showing the possible outcomes of combining these sufficient when analyzing the nomothetic relation- variables. ships among variables. It is not uncommon, how- ever, to find causal factors that are either 100 A suffident cause, on the other hand, represents percent necessary (you must be female to become a conclition that, if it is present, guarantees the ef- pregnant) or 100 percent sufficient (skipping an fect in question. This is not to say that a sufficient exam will inevitably cause you to fail it). In the idiographic analysis of single cases, you may reach a depth of explanation from which it is

94 Chapter 4: Research Design FF guish betvveen the unit of analysis and the aggre- gates that we generalize aboue For instance, a FF F F researcher may study a class of people, such as F FF Democrats, college undergraduates, African Ameri- can women under 30, or some other collection. But F FF if the researcher is interested in exploring, descTib- ing, or explaining how different groups of individu- ADD A als behave as individuals, the unit of analysis is the individual, not the group. This is true even though BB c the researcher uses the information about individu- als to generalize about aggregates of individuals, as B C CA in saying that more Democrats than Republicans favor legalizing marijuana, Think of it this way: C Bo 0 Having an attitude about marijuana is something that can only be an attribute of an individual, not a A group; that is, there is no one group \"mind\" that CC A can have an attitudeo So even when we generalize about Democrats, we're generalizing about an at- FIGURE 4-3 tribute they possess as individuals, Sufficient Causeo Not taking the exam is asufficient cause of failing it, even though there are other ways of failing (such as In contrast, we may sometimes want to study answering randomly)o groups, considered as individual \"actors\" or entities that have attributes as gTOUpS, For instance, we reasonable to assume that things could not have might want to compare the characteristics of differ- turned out differently, suggesting you have deter- ent types of street gangso In that case our unit of mined the sufficient causes for a particular result. analysis would be gangs (not members of gangs), (Anyone with all the same details of your genetic and we might proceed to make generalizations inheritance, upbringing, and subsequent experi- about different types of gangs. For example, we ences would have ended up going to college,) At might conclude that male gangs are more violent the same tin1e, there could always be other causal than female gangso Each gang (unit of analysis) paths to the same resulL Thus, the idiographic would be described in terms of two variables: causes are sufficient but not necessary. (1) What sex are the members? and (2) How vio- lent are its activities? So we might study 52 gangs, Units of Analysis reporting that 40 were male and 12 were female, and so forth. The \"gang\" would be the unit of In social research, there is virtually no limit to what analysis, even though some of the characteristics or whom can be studied, or the units of analysis. were drawn from the components (members) of This topic is relevant to all forms of social research, although its implications are clearest in the case of the gangs. nomothetic, quantitative studies, Social researchers tend to choose individual The idea for units of analysis may seem slippery people as their units of analysiso You may note the at first, because research-especially nomothetic characteristics of individual people-gender, age, research- often studies large collections of people region of birth, attitudes, and so fortho You can then or things, or aggregateso It's important to distin- combine these descriptions to provide a composite picture of the group the individuals represent, units of analysis The what or whom being whether a street-corner gang or a whole societyo studied. In social science research, the most typical units of analysis are individual people. For example, you may note the age and gender of each student enrolled in Political Science 110 and then characterize the group of students as

Units of Analysis 95 being 53 percent men and 47 percent women and analysis. For example, we may want to discover as having a mean age of 1806 yearso Although the what kinds of college students (individuals) are final description would be of the class as a whole, most successful in their careers; we may also want the description is based on characteristics that to learn what kinds of colleges (organizations) pro- members of the class have as individuals, duce the most successful graduateso The same distinction between units of analysis Here's an example that illustrates the complex- and aggregates occurs in explanatory studieso ity of units of analysis. Murder is a fairly personal Suppose you wished to discover whether students matter: One individual kills another individuaL with good study habits received better grades in However, when Charis Kubrin and Ronald Weitzer political Science 110 than students with poor study (2003: 157) ask, \"Why do these neighborhoods habits did. You would operationalize the variable generate high homicide rates?\" the unit of analysis study habits and measure this variable, perhaps in in that phrase is neighborhood. You can probably terms of hours of study per week, You might then imagine some kinds of neighborhoods (eogo, poor, aggregate students with good study habits and urban) that would have high homicide rates and those with poor study habits and see which some (e.g., wealthy, suburban) that would have group received the best grades in the course. The low rates. In this particular conversation, the unit purpose of the study would be to explain why of analysis (neighborhood) would be categorized in some groups of students do better in the course terms of variables such as economic level, locale, and than others do, but the unit of analysis is still indi- homicide rate, vidual students. In their analysis, however, Kubrin and Weitzer Units of analysis in a study are usually also the were also interested in different types of homicide: units of observation, Thus, to study success in a po- in particular, those that occurred in retaliation for litical science course, we would observe individual some earlier event, such as an assault or insulL Can students. Sometimes, however, we \"observe\" our you identify the unit of analysis common to all of units of analysis indirectly. For example, suppose the following excerpts? we want to find out whether disagreements about the death penalty tend to cause divorce. In this 10 The sample of killings 000 case, we might \"observe\" individual husbands and wives by asking them about their attitudes about 20 The coding instrument includes over 80 items capital punishment. in order to distinguish couples related to the homicide, who agree and disagree on this issue. In this case, our units of observation are individual wives and 3.. Of the 2,161 homicides that occurred from husbands, but our units of analysis (the t11ings we 1985 [tol1995 0,0 want to study) are couples. 4. Of those with an identified motive, 19.5 Units of analysis, then, are those things we ex- percent (n = 337) are retaliatoryo amine in order to create summary descriptions of all such units and to explain differences among (Kubrin and Weitzer 2003: 163) them. In most research projects, the unit of analysis will probably be clear to YOUo When the unit of In each of these excerpts, the unit of analysis is analysis is not clear, however, it's essential to deter- homicide (also called killing or murder). Sometimes mine what it is; otherwise, you cannot determme you can identify the unit of analysis in the descrip- what observations are to be made about whom tion of the sampling methods, as in the first ex- or what. cerpt. A discussion of classification methods might also identify the unit of analysis, as in the second Some studies try to describe or explain more excerpt (80 ways to code the homicides). Often, than one unit of analysis, In these cases, the re- numerical summaries point the way: 2,161 homi- searcher must anticipate what conclusions she or cides; 19.5 percent (of the homicides). With a little he wishes to draw with regard to which units of practice you'll be able to identify the units of analy- sis in most social research reports, even when more than one is used in a given analysiso

96 Chapter 4: Research Design To explore this topic in more depth, let's con- aggregate these individuals and make generaliza- sider several common units of analysis in social tions about the population they belong to. research. Groups Individuals Social groups can also be units of analysis in social As mentioned, individual human beings are per- research. That is, we may be interested in charac- haps the most typical units of analysis for social re- teristics that belong to one group, considered as a search. Social researchers tend to describe and ex- single entity. If you were to study the members of a plain social groups and interactions by aggregating criminal gang to learn about criminals, the individ- and manipulating the descriptions of individuals. ual (criminal) would be the unit of analysis; but if you studied all the gangs in a city to learn the dif- Any type of individual may be the unit of ferences, say, between big gangs and small ones, analysis for social research. This point is more im- between \"uptown\" and \"downtown\" gangs, and so portant than it may seem at first. The norm of gen- forth, you would be interested in gangs rather than eralized understanding in social research should their individual members. In this case, the unit of suggest that scientific findings are most valuable analysis would be the gang, a social group. when they apply to all kinds of people. In practice, however, social researchers seldom study all kinds Here's another example. Suppose you were in- of people. At the very least. their studies are typi- terested in the question of access to computers in cally limited to the people living in a single country, different segments of society. You might describe though some comparative studies stretch across na- families in terms of total annual income and ac- tional boundaries. Often, though, studies are quite cording to whether or not they had computers. You circumscribed. could then aggregate families and describe the mean income of families and the percentage with Examples of classes of individuals that might be computers. You would then be in a position to de- chosen for study include students, gays and les- tennine whether families with higher incomes bians, auto workers, voters, single parents, and fac- were more likely to have computers than were ulty members. Note that each of these terms im- those with lower incomes. In this case, the unit plies some population of individuals. Descriptive of analysis would be families. studies with individuals as their units of analysis typically aim to describe the population that com- As with other units of analysis, we can derive prises those individuals, whereas explanatory stud- the characteristics of social groups from those of ies aim to discover the social dynamics operating their individual members. Thus, we might describe within that population. a family in terms of the age, race, or education of its head. In a descriptive study, we might find the As the units of analysis, individuals may be percentage of all families that have a college- characterized in terms of their membership in so- educated head of family. In an explanatory study, cial groupings. Thus, an individual may be de- we might determine whether such families have, scribed as belonging to a rich family or to a poor on average, more or fewer children than do families one, or a person may be described as having a headed by people who have not graduated from college-educated mother or not. We might examine college. In each of these examples, the family is the in a research project whether people with college- unit of analysis. In contrast, had we asked whether educated mothers are more likely to attend college college-educated individuals have more or fewer than are those with non-college-educated mothers children than do their less-educated counterparts, or whether high school graduates in rich families then the individual would have been the unit of are more likely than those in poor families to at- analysis. tend college. In each case, the unit of analysis-the \"thing\" whose characteristics we are seeking to Other units of analysis at the group level could describe or explain-is the individual. We then be friendship cliques, married couples, census

Units of Analysis 97 blocks, cities, or geographic regions. As with indi- Even though individuals are usually the actors viduals, each of these terms implies some popula- in social interactions, there is a difference between tion. Street gangs implies some population that in- (1) comparing the kinds of people who subscribe cludes all street gangs, perhaps in a given city. You to different Internet service providers (individuals might then describe this population by generalizing being the unit of analysis) and (2) comparing the from your findings about individual gangs. For in- length of chat-room discussions on those same ISPs stance, you might describe the geographic distribu- (the discussion being the unit of analysis). tion of gangs throughout a city. In an explanatory study of street gangs, you might discover whether Social Artifacts large gangs are more likely than small ones to en- gage in intergang warfare. Thus, you would arrive Another unit of analysis is the social artifact or at conclusions about the population of gangs by us- any product of social beings or their behavior. One ing individual groups as your unit of analysis. class of artifacts includes concrete objects such as books, poems, paintings, automobiles, bUildings, Organizations songs, pottery, jokes, student excuses for missing exams, and scientific discoveries. Formal social organizations may also be the units of analysis in social research. For example, a re- For example, Lenore Weitzman and her associ- searcher might study corporations, by which he or ates (1972) were interested in learning how gender she implies a population of all corporations. Indi- roles are taught. They chose children's picture vidual corporations might be characterized in terms books as their unit of analysis. Specifically, they ex- of their number of employees, net annual profits, amined books that had received the Caldecott gross assets, number of defense contracts, percent- Medal. Their results were as follows: age of employees from racial or ethnic minority groups, and so forth. We might determine whether We found that females were underrepresented large corporations hire a larger or smaller percent- in the titles, central roles, pictures, and stories age of minority group employees than do small of every sample of books we examined. Most corporations. Other examples of formal social or- children's books are about boys, men, male ani- ganizations suitable as units of analysis include mals, and deal exclusively with male adven- church congregations, colleges, army divisions, aca- tures. Most pictures show men singly or in demic departments, and supermarkets. groups. Even when women can be found in the books, they often play insignificant roles, re- Figure 4-4 provides a graphic illustration of maining both inconspicuous and nameless. some different units of analysis and the statements that might be made about them. (Weitzman et al.. 1972: 1128) Social Interactions In a more recent study, Roger Clark, Rachel Lennon, and Leana Morris (1993) concluded that Sometimes social interactions are the relevant units male and female characters are now portrayed less of analysis. Instead of individual humans, you can stereotypically than before, observing a clear study what goes on between them: telephone calls, progress toward portraying men and women in kisses, dancing, arguments, fistfights, e-mail ex- nontraditional roles. However, they did not find to- changes, chat-room discussions, and so forth. As tal equality between the sexes. you saw in Chapter 2, social interaction is the basis for one of the primary theoretical paradigms in the As this example suggests, just as people or so- social sciences, and the number of units of analysis cial groups imply popUlations, each social object that social interactions provide is nearly infinite. social artifact Any product of social beings or their behavior. Can be a unit of analysis.

60% of the sample are women 10% of the sample are wearing an eye patch 10% of the sample have pigtails 20% of the families have a single parent 50% of the families have two children 20% of the famillies have no children The mean number of children per family is 1. 3 20% of the households are occupied by more than one family 30% of the households have holes in their roofs 10% of the households are occupied by aliens Notice also that 33% of the families live in multiple-family households with family as the unit of analysis FIGURE 4-4 Illustrations of Units of Analysis. Units of analysis in social research can be individuals, groups, or even nonhuman entities.

Units of Analysis 99 implies a set of all objects of the same class: all the individual chairpersons-would be our units of books, all novels, all biographies, all introductory analysis. sociology textbooks, all cookbooks, all press conferences. In a study using books as the units of Units ofAna/ysis in Review analysis, an individual book might be characterized by its size, weight, length, price, content, number The examples in this section should suggest the of pictures, number sold, or description of the au- nearly infinite variety of possible units of analysis thor. Then the population of all books or of a par- in social research. Although individual human be- ticular kind of book could be analyzed for the ings are typical objects of study. many research purpose of description or explanation: what kinds questions can be answered more appropriately of books sell best and why, for example. through the examination of other units of analysis. Indeed, social researchers can study just about any- Similarly, a social researcher could analyze thing that bears on social life. whether paintings by Russian, Chinese, or US. artists showed the greatest degree of working-class Moreover, the types of units of analysis named consciousness, taking paintings as the units of in this section do not begin to exhaust the possibili- analysis and describing each, in part, by the nation- ties. Morris Rosenberg (1968: 234-48), for ex- ality of its creator. Or you might examine a news- ample, speaks of individual, group, organizational, paper's editorials regarding a local university, for institutional, spatial, cultural, and societal units of the purpose of describing, or perhaps explaining, analysis. Jolm and Lyn Lofland (1995: 103-13) changes in the newspaper's editorial position on speak of practices, episodes, encounters, roles, rela- the university over tin1e. In this example, individ- tionships, groups, organizations, settlements, social ual editorials would be the units of analYSis. worlds, lifestyles, and subcultures as suitable units of study. The important thing here is to grasp the Social interactions form another class of social logic of units of analysis. Once you do, the possibili- artifacts suitable for social research. For example, ties for fruitful research are limited only by your we might characterize weddings as racially or reli- imagination. giously mixed or not, as religious or secular in cere- mony, as resulting in divorce or not, or by descrip- Categorizing possible units of analysis might tions of one or both of the marriage partners (such make the concept seem more complicated than it as \"previously married,\" \"Oakland Raider fan,\" needs to be. What you call a given unit of analy- \"wanted by the FBI\"). When a researcher reports sis-a group, a formal organization, or a social arti- that weddings between partners of different reli- fact-is irrelevant The key is to be clear about gions are more likely to be performed by secular what your unit of analysis is. When you embark on authorities than those between partners of the a research project, you must decide whether you're same religion are, the weddings are the units of studying marriages or marriage partners, crin1es or analysis, not the individuals involved. criminals, corporations or corporate executives. Otherwise, you run the risk of drawing invalid con- Other social interactions that might be units of clusions because your assertions about one unit of analysis are friendship choices, court cases, traffic analysis are actually based on the examination of aCcidents, divorces, fistfights, ship launchings, air- another. We'll see an example of this issue in the line hijackings, race riots, final exams, student next section as we look at the ecological fallacy. demonstrations, and congressional hearings. Con- gressional hearings, for instance, could be charac- Faulty Reasoning about Units terized by whether or not they occurred during an ofAnalysis: The Ecological Fallacy election campaign, whether the committee chairs and Reductionism were running for a higher office, whether they had received campaigns contributions from interested At this point, it's appropriate to introduce two parties, and so on. Notice that even if we character- types of faulty reasoning that you should be aware ized and compared the hearings in terms of the of: the ecological fallacy and reductionism. Each committee chairs, the hearings themselves-not

100 Chapter 4: Research Design represents a potential pitfall regarding units of not know for sure that more Protestants than analysis, and either can occur in doing research Catholics committed suicide. and drawing conclusions from the results. In spite of these hazards, social researchers of- The Ecological Fallacy ten have little choice but to address a particular research question through an ecological analysis. In this context, \"ecological\" refers to groups or Perhaps the most appropriate data are simply not sets or systems: something larger than individuals. available. For example, the precinct vote tallies and The ecological fallacy is the assumption that the precinct characteristics mentioned in our initial something learned about an ecological unit says example may be easy to obtain, but we may not something about the individuals making up that have the resources to conduct a postelection survey unit. Let's consider a hypothetical illustration of this of individual voters. In such cases, we may reach a fallacy. tentative conclusion, recognizing and noting the risk of an ecological fallacy. Suppose we're interested in learning something about the nature of electoral support received by a Although you should be careful not to commit female political candidate in a recent citywide elec- the ecological fallacy, don't let these warnings lead tion. Let's assume we have the vote tally for each you into committing what we might call the indi- precinct so we can tell which precincts gave her the vidualistic fallacy Some people who approach so- greatest support and which the least. Assume also cial research for the first time have trouble recon- that we have census data describing some charac- ciling general patterns of attitudes and actions with teristics of these precincts. Our analysis of such data individual exceptions. But generalizations and might show that precincts with relatively young probabilistic statements are not invalidated by indi- voters gave the female candidate a greater propor- vidual exceptions. Your knovving a rich Democrat, tion of their votes than did precincts with older vot- for example, doesn't deny the fact that most rich ers. We might be tempted to conclude from these people vote Republican-as a general pattern. Sim- findings that younger voters are more likely to ilarly, if you know someone who has gotten rich vote for female candidates than older voters are- without any formal education, that doesn't deny in other words, that age affects support for the the general pattern of higher education relating to woman. In reaching such a conclusion, we run higher income. the risk of committing the ecological fallacy, because it may have been the older voters in those The ecological fallacy deals with something else \"young\" precincts who voted for the woman. Our altogether-confusing units of analysis in such a problem is that we have examined precil1cts as our way that we draw conclusions about individuals units of analysis but wish to draw conclusions solely from the observation of groups. Although about voters. the patterns observed between variables at the level of groups may be genuine, the danger lies in rea- The same problem would arise if we discovered soning from the observed attributes of groups to that crime rates were higher in cities having large the attributes of the individuals who made up African American populations than in those with those groups, even though we have not actually few African Americans. We would not know if the observed individuals. crimes were actually committed by African Ameri- cans. Or if we found suicide rates higher in Protes- Reductionism tant countries than in Catholic ones, we still could A second type of potentially faulty reasoning re- ecological fallacy Erroneously drawing conclu- lated to units of analysis is reductionism. Reduc- sions about individuals solely from the observation tionism involves attempts to explain a particular of groups. phenomenon in terms of limited and/or lower- order concepts. The reductionist explanation is not altogether wrong; it is simply too limited. Thus, you

The TIme Dimension 101 might attempt to predict this year's winners and nomic plight of the colonies in relation to Britain? losers in the National Basketball Association by fo- The megalomania of the founders? As soon as we cusing on the abilities of the individual players on inquire about the single cause, we run the risk of each team. This is certainly not stupid or irrelevant, reductionism. If we were to regard shared values as but the success or failure of teams involves more the cause of the American Revolution, our unit of than just the individuals in them; it involves coach- analysis would be the individual colonist. An econ- ing, teamwork strategies, finances, facilities, fan omist, though, might choose the 13 colonies as loyalty, and so forth. To understand why some units of analYSis and examine the economic organi- teams do better than others, you would make zations and conditions of each. A psychologist \"team\" the unit of analysis, and the quality of might choose individual leaders as the units of players would be one variable you would prob- analysis for purposes of examining their personali- ably want to use in describing and classifying the ties. Of course, there's nothing wrong in choosing teams. these units of analysis as pan of an ex'Planation of the An1erican Revolution, but I think you can Further, different academic disciplines ap- see how each alone would not produce a complete proach the same phenomenon quite differently. answer. Sociologists tend to consider sociological variables (such as values, norms, and roles), economists Like the ecological fallacy, reductionism can ponder economic variables (such as supply and occur when we use inappropriate units of analysis. demand and marginal value), and psychologists The appropriate unit of analysis for a given re- examine psychological variables (such as person- search question, however, is not always clear. ality types and traumas). Explaining all or most Social researchers, especially across disciplinary human behavior in terms of economic factors is boundaries, often debate this issue. called economic reductionism, explaining it in terms of psychological factors is called psychologi- The Time Dimension cal reductionism, and so forth. Notice how this is- sue relates to the discussion of theoretical para- So far in this chapter, we've regarded research de- digms in Chapter 2. sign as a process for deciding what aspects we'll ob- serve, of whom, and for what purpose. Now we For many social scientists, the field of socio- must consider a set of time-related options that cuts biology is a prime example of reductionism, sug- across each of these earlier considerations. We can gesting that all social phenomena can be explained choose to make observations more or less at one in terms of biological factors. Thus, for example, time or over a long period. Edward O. Wilson (1975) sought to explain altruis- tic behavior in human beings in terms of genetic Tin1e plays many roles in the design and execu- makeup. In his neo-Darwinian view, Wilson sug- tion of research, quite aside from the time it takes gests that humans have evolved in such a way that to do research. Earlier we noted that the time individuals sometimes need to sacrifice themselves sequence of events and situations is critical to for the benefit of the whole species. Some people might ex'Plain such sacrifice in terms of ideals or reductionism A fault of some researchers: a strict warm feelings between humans. However, genes lin1itation (reduction) of the kinds of concepts to be are the essential unit in Wilson's paradigm, produc- considered relevant to the phenomenon under ing his famous dictum that human beings are \"only study. DNA's way of making more DNA.\" SOCiobiology A paradigm based in the view that Reductionism of any type tends to suggest that social behavior can be eX\"plained solely in terms of particular units of analysis or variables are more genetic characteristics and behavior. relevant than others. Suppose we ask what caused the American Revolution. Was it a shared commit- ment to the value of individual liberty? The eco-

102 Chapter 4: Research Design determining causation (a point we'll return to in brought about by a series of national campaigns, Part 4). Time also affects the generalizability of such as the Great Proletarian Cultural Revolution, research findings. Do the descriptions and explana- dating from the Chinese Revolution in 1949 (which tions resulting from a particular study accurately brought the Chinese Communists into power) and represent the situation of ten years ago, ten years continuing into the present. from now, or only the present? Researchers have two principal options available to deal with the is- These campaigns altered political atmospheres sue of time in the design of their research: cross- and affected people's work and nonwork activi- sectional studies and longitudinal studies. ties. Because of these campaigns, it is difficult to draw conclusions from a cross-sectional social Cross-Sectional Studies survey, such as the one presented in this book, about general patterns of Chinese workplaces A cross-sectional study involves observations of a and their effects on workers. Such conclusions sample, or cross section, of a population or phe- may be limited to one period of time and are nomenon that are made at one point in time. Ex- subject to further tests based on data collected ploratory and descriptive studies are often cross- at other times. sectionaL A single US. Census, for instance, is a study aimed at describing the US. population at a (1994,' 19) given time. The problem of generalizations about social Many eX'Planatory studies are also cross- life from a \"snapshot\" is one this book repeatedly sectionaL A researcher conducting a large-scale na- addresses. One solution is suggested by Bian's fi- tional survey to examine the sources of racial and nal comment-about data collected \"at other religious prejudice would, in all likelihood, be deal- times\": Social research often involves revisiting ing with a single time frame-taking a snapshot so phenomena and building on the results of earlier to speak, of the sources of prejudice at a particular research. point in history. Longitudinal Studies Explanatory cross-sectional studies have an inherent problem. Although their conclusions In contrast to cross-sectional studies, a longitudi- are based on observations made at only one time, nal study is designed to permit observations of the typically they aim at understanding causal pro- same phenomenon over an extended period. For cesses that occur over time. This problem is some- example, a researcher can participate in and ob- what akin to that of determining the speed of a serve the activities of a UFO cult from its inception moving object on the basis of a high-speed, still to its demise. Other longitudinal studies use records photograph that freezes the movement of the or artifacts to study changes over tinle. In analyses object. of newspaper editorials or Supreme Court decisions over time, for example, the studies are longitudinal Yanjie Bian, for example, conducted a survey of whether the researcher's actual observations and workers in Tianjin, China, for the purpose of study- analyses were made at one time or over the course ing stratification in contemporary, urban Chinese of the actual events under study. society. In undertaking the survey in 1988, how- ever, he was conscious of the important changes Many field research projects, involving direct observation and perhaps in-depth interviews, are cross-sectional study A study based on observa- naturally longitudinaL Thus, for example, when tions representing a single point in time. Ramona Asher and Gary Fine (1991) studied the longitudinal study A study design involving the life experiences of the ,vives of alcoholic men, they collection of data at different points in time., were in a position to examine the evolution of troubled marital relationships over time, sometimes

The Time Dimension 103 even including the reactions of the subjects to the An in-depth analysis, however, indicates that research itself. the slight increase in political knowledge resulted from the fact that the people in the 1989 sample In the classic study When Prophecy Fails (1956), were more highly educated than those from earlier Leon Festinger, Henry Reicker, and Stanley samples were. When educational levels were taken Schachter were specifically interested in learning into account, the researchers concluded that politi- what happened to a flying saucer cult when their cal knowledge has actually declined within specific predictions of an alien encounter failed to come educational groups. true. Would the cult members close down the group, or would they become all the more commit- Cohort Studies ted to their beliefs? A longitudinal study was re- quired to provide an answer. (They redoubled their In a cohort study, a researcher examines specific efforts to get new members.) subpopulations, or cohorts, as they change over time, Typically, a cohort is an age group, such as Longitudinal studies can be more difficult for people born during the 1950s, but it can also be quantitative studies such as large-scale surveys. some other time grouping, such as people born Nonetheless, they are often the best way to study during the Vietnam War, people who got married changes over time. There are three special types in 1994, and so forth. An example of a cohort of longitudinal studies that you should know study would be a series of national surveys, con- about: trend studies, cohort studies, and panel ducted perhaps every 20 years, to study the atti- studies. tudes of the cohort born during World War II to- ward US. involvement in global affairs. A sample Trend Studies of people 15-20 years old might be surveyed in 1960, another sample of those 35-40 years old in A trend study is a type of longitudinal study that 1980, and another sample of those 55-60 years examines changes within a population over time. A old in 2000. Although the specific set of people simple example is a comparison of US. Censuses studied in each survey would differ, each sample over a period of decades, showing shifts in the would represent the cohort born between 1940 makeup of the national population. A similar use of and 1945. archival data was made by Michael Carpini and Scott Keeter (1991), who wanted to know whether James Davis (1992) turned to a cohort analysis contemporary US. citizens were better or more in an attempt to understand shifting political orien- poorly informed about politics than citizens of an tations during the 1970s and 1980s in the United earlier generation were. To find out, they compared States. OveralL he found a liberal trend on issues the results of several Gallup Polls conducted during the 1940s and 1950s 'with a 1989 survey that asked trend study A type of longitudinal study in which several of the same questions tapping political a given characteristic of some population is moni- knowledge. tored over time. An example would be the series of Gallup Polls showing the electorate's preferences for OveralL the analysis suggested that contempo- political candidates over the course of a campaign, rary citizens were slightly better informed than ear- even though different samples were interviewed at lier generations were. In 1989, 74 percent of the each point., sample could name the vice president of the United States, compared with 67 percent in 1952. Sub- cohort study A study in which some specific sub- stantially higher percentages of people in 1989 popUlation, or cohort, is studied over time, although than in 1947 could explain presidential vetoes and data may be collected from different members in congressional overrides of vetoes. On the other each set of observations, For example, a study of the hand, more of the 1947 sample could identify their occupational history of the class of 1970 in which US. representative (38 percent) than the 1989 questionnaires were sent every five years would be a sample (29 percent) could. cohort study,

104 Chapter 4: Research Design TABLE 4-1 and Political Liberalism Survey dates 1972 to 1974 1977to 1980 1982 to 1984 1987to 1989 of cohort 20-24 25-29 30-34 35-39 72% Percent who would let 68% 73% 73% the Communist speak such as race, gender, religion, politics, crime, and them to reduce their commitment to educational free speech, But did this trend represent people in funding. Moreover, cross-sectional data support general getting a bit more liberal, or did it merely that expectation, The researchers present several reflect liberal younger generations replacing the data sets showing those over 65 voicing less support conservative older ones? for education funding than those under 65 did. To answer this question, Davis examined na- Such simplistic analyses, however, leave out an tional surveys (from the General Social Survey, of important variable: increasing support for educa- which he is a founder) conducted in four time peri- tional funding in US. society over time in general, ods, five years aparL In each survey, he grouped The researchers add to this the concept of \"genera- the respondents into age groups, also five years tional replacement,\" meaning that the older re- apart. This strategy allowed him to compare differ- spondents in a survey grew up during a time when ent age groups at any given point in time as well as there was less support for education in generaL follow the political development of each age group whereas the younger respondents grew up during over time. a time of greater overall support. One of the questions he examined was A cohort analysis allowed the researchers to whether a person who admitted to being a Com- determine what happened to the attitudes of munist should be allowed to speak in the respon- specific cohorts over time. Here, for example, are dents' communities. Consistently, the younger re- the percentages of Americans born during the spondents in each time period were more ,>\\Tilling to 1940s who felt educational spending was too low, let the Communist speak than the older ones were. when members of that cohort were interviewed Among those aged 20-40 in the first set of the sur- over time (Plutzer and Berkman, 2000: 76): vey, for example, 72 percent took this liberal posi- tion, contrasted with 27 percent among respon- Year Interviewed Percent Who Say Educational dents 80 and oldeL What Davis found when he 1970s examined the youngest cohort over time is shown 19805 Too Low in Table 4-1. This pattern of a slight, conservative 19905 shift in the 1970s, followed by a liberal rebound in 20005 58 the 1980s, typifies the several cohorts Davis ana- 66 lyzed (J. Davis 1992: 269), 74 79 In another study, Eric Plutzer and Michael Berkman (2005) used a cohort design to completely As these data indicate, those who were born reverse a prior conclusion regarding aging and sup- during the 1940s have steadily increased their sup- port for education. Logically, as people grow well port for educational funding as they have passed beyond the child-rearing years, we might expect through and beyond the child-rearing years. panel study A type of longitudinal study, in which Panel Studies data are collected from the same set of people (the sample or panel) at several points in time. Though similar to trend and cohort studies, a panel study examines the same set of people each

TIle Time Dimension 105 time. For example, we could interview the same affiliation. A trend study might look at shifts in US. sample of voters every month during an election religious affiliations over time, as the Gallup Poll campaign, asking for whom they intended to vote. does on a regular basis, A cohort study might fol- Though such a study would allow us to analyze low shifts in religious affiliations among \"the De- overall trends in voter preferences for different can- pression generation,\" specifically, say, people who didates, it would also show the precise patterns of were 20 to 30 years old in 1932. We could study a persistence and change in intentions. For example, sample of people 30-40 years old in 1942, a new a trend study that showed that Candidates A and B sample of people aged 40-50 in 1952, and so forth. each had exactly half of the voters on September 1 A panel study could start with a sample of the and on October 1 as well could indicate that none whole population or of some special subset and of the electorate had changed voting plans, that all study those specific individuals over time. Notice of the voters had changed their intentions, or that only the panel study would give a full picture something in-between. A panel study would elimi- of the shifts among the various categories of affilia- nate this confusion by showing what kinds of vot- tions, including \"none.\" Cohort and trend studies ers switched from A to B and what kinds switched would uncover only net changes. from B to A, as well as other facts. Longitudinal studies have an obvious advan- Joseph Veroff, Shirley Hatchett, and Elizabeth tage over cross-sectional ones in providing informa- Douvan (1992) wanted to learn about marital ad- tion describing processes over time. But this advan- justment among newlyweds, specifically regarding tage often comes at a heavy cost in both time and differences between white and African American money, especially in a large-scale survey. Observa- couples. To get subjects for study, they selected a tions may have to be made at the tin1e events are sample of couples who applied for marriage li- occurring, and the method of observation may re- censes in Wayne County, Michigan, April through quire many research workers. June 1986. Panel studies, which offer the most compre- Concerned about the possible impact their re- hensive data on changes over time, face a special search might have on the couples' marital adjust- problem: panel attrition. Some of the respondents ment, the researchers divided their sample in half studied in the first wave of the survey might not at random: an experimental group and a control participate in later waves. (This is comparable to group (concepts we'll explore further in Chapter the problem of experimental mortality discussed in 8). Couples in the former group were intensively Chapter 8.) The danger is that those who drop out interviewed over a four-year period, whereas the of the study may be atypicaL thereby distorting the latter group was contacted only briefly each yeaL results of the study. Thus, when Carol Aneshensel and her colleagues conducted a panel study of ado- By studying the same couples over time, the re- lescent girls (comparing Latinas and non-Latinas), searchers could follow the specific problems that they looked for and found differences in character- arose and the way the couples dealt with them. As istics of survey dropouts among Latinas born in a by-product of their research, they found that the United States and those born in Mexico. These those studied the most intensely seemed to achieve differences needed to be taken into account to a somewhat better marital adjustment. The re- avoid misleading conclusions about differences searchers felt that the interviews could have forced between Latinas and non-Latinas (Aneshensel couples to discuss matters they might have other- et aL 1989), wise buried. Comparing the Three Types Approximating Longitudinal Studies ofLongitudinal Studies Longitudinal studies do not always provide a feasi- To reinforce the distinctions among trend, cohort, ble or practical means of studying processes that and panel studies, let's contrast the three study take place over time, Fortunately, researchers often designs in terms of the same variable: religiolls can draw approxin1ate conclusions about such

106 Chapter 4: Research Design processes even when only cross-sectional data are pressure, and so forth. By reading across the age- available. Here are some ways to do that. group ratings for each health condition, you would have something approximating the health history Sometimes cross-sectional data imply processes of individuals. Thus, you might conclude that the over time on the basis of simple logic For example, average person develops vision problems before in the study of student drug use conducted at the hearing problems. You would need to be cautious in University of Hawaii (Chapter 2), students were this assumption, however. because the differences asked to report whether they had ever tried each of might reflect societyvvide trends. Perhaps improved several illegal drugs . The study found that some hearing examinations instituted in the schools had students had tried both marijuana and LSD, some affected only the young people in your study. had tried only one, and others had tried neither. Because these data were collected at one time, and Asking people to recall their pasts is another because some students presumably would experi- common way of approximating observations over ment with drugs later on, it would appear that such time. Researchers use that method when they ask a study could not tell whether students were more people where they were born or when they gradu- likely to try marijuana or LSD first. ated from high school or whom they voted for in 1988. Qualitative researchers often conduct in- A closer examination of the data showed, how- depth \"life history\" interviews. For example, ever, that although some students reported having C. Lynn Carr (1998) used this technique in a study tried marijuana but not LSD, there were no stu- of \"tomboyism.\" Her respondents, aged 25-40, dents in the study who had tried only LSD. From were asked to reconstruct aspects of their lives this finding it was inferred-as common sense from childhood on, including experiences of identi- suggested-that marijuana use preceded LSD fying themselves as tomboys. use. If the process of drug experimentation oc- curred in the opposite time order, then a study at a The danger in this technique is evident. Some- given time should have found some students who times people have faulty memories; sometimes had tried LSD but not marijuana, and it should they lie. When people are asked in postelection have found no students who had tried only polls whom they voted for. the results inevitably marijuana. show more people voting for the winner than actu- ally did so on election day. As part of a series of in- Researchers can also make logical inferences depth interviews, such a report can be validated in whenever the time order of variables is clear. If we the context of other reported details; however, re- discovered in a cross-sectional study of college stu- sults based on a single question in a survey must be dents that those educated in private high schools regarded with caution. received better college grades than those educated in public high schools did, we would conclude that This discussion of the ways that time figures the type of high school attended affected college into social research suggests several questions you grades, not the other way around. Thus, even should confront in your own research projects. In though we made our observations at only one designing any study, be sure to look at both the time, we would feel justified in drawing conclu- e;.,:plicit and implicit assumptions you're making sions about processes taking place across time. about time. Are you interested in describing some process that occurs over time, or are you simply go- Very often, age differences discovered in a ing to describe what exists now? If you want to de- cross-sectional study form the basis for inferring scribe a process occurring over time, ,viII you be processes across time. Suppose you're interested in able to make observations at different points in the the pattern of worsening health over the course of process, or will you have to approximate such ob- the typical life cycle. You might study the results of servations by drawing logical inferences from what annual checkups in a large hospitaL You could you can observe now? If you opt for a longitudinal group health records according to the ages of those design, which method best serves your research examined and rate each age group in terms of purposes? several health conditions-sight, hearing, blood

How to Design a Research Project 107 Examples ofResearch Strategies I> Using interview and observational field data, I demonstrate how a system of temporary As the preceding discussions have implied, social employment in a participative workplace both research follows many paths. The following short exploited and shaped entry-level workers' excerpts further illustrate this point. As you read aspirations and occupational goals, (V. Smith each eXCel})!, note both the content of each study 1998:411) and the method used to study the chosen topic Does the study seem to be exploring, describing, or I> I collected data [on White Separatist rhetoric] explaining (or some combination of these)? What from several media of public discourse, includ- are the sources of data in each study? Can you ing periodicals, books, pamphlets, transcripts identify the unit of analysis? Is the dimension of from radio and television talk shows, and time relevant? If so, how\"vill it be handled? newspaper and magazine accounts. (Berbrier 1998: 435) I> This case study of unobtrusive mobilizing by Southern California Rape Crisis Center uses ar- I> In the analysis that follows, racial and gender chivaL observational, and interview data to ex- inequality in employment and retirement will plore how a feminist organization worked to be analyzed, using a national sample of persons change police, schools, prosecutors, and some who began receiving Social Security Old Age state and national organizations from 1974 to benefits in 1980-8L (Hogan and Perrucci 1994. (Schmitt and Martin 1999: 364) 1998: 528) I> Using life history narratives, the present study I> Drawing from interviews with female crack investigates processes of agency and conscious- dealers, this paper explores the techniques ness among 14 women who identified them- they use to avoid arrest (Jacobs and Miller selves as tomboys. (Carr 1998: 528) 1998: 550) I> By drawing on interviews with activists in the How to DeSign former Estonian Soviet Socialist Republic, we a Research Project specify the conditions by which accommoda- tive and oppositional subcultures exist and are You've now seen some of the options available to successfully transformed into social move- social researchers in designing projects. I know ments. (Johnston and Snow 1998: 473) there are a lot of components, and the relationships among them may not be totally clear, so here's a I> This paper presents the results of an ethno- way of pulling them together. Let's assume you graphic study of an AIDS service organization were to undertake research, Where would you located in a small city It is based on a combina- start? Then, where would you go? tion of participant observation, interviews with participants, and review of organizational Although research design occurs at the begin- records. (Kilburn 1998: 89) ning of a research project, it involves all the steps of the subsequent project. This discussion, then, pro- I> Using interviews obtained during fieldwork in vides both guidance on how to start a research Palestine in 1992, 1993, and 1994, and employ- project and an overview of the topics that follow in ing historical and archival records, I argue that later chapters of this book Palestinian feminist discourses were shaped and influenced by the sociopolitical context in Figure 4-5 presents a schematic view of the so- which Palestinian women acted and with cial research process. I present this view reluctantly, which they interacted. (Abdulhadi 1998: 649) because it may suggest more of a step-by-step order to research than actual practice bears out. I> This article reports on women's experiences of Nonetheless, this idealized overview of the process breastfeeding in public as revealed through in- depth interviews with 51 women. (Stearns 1999: 308)

X~Y A_B_E_F A~B ~;t ;t~ ~ C-D-X-Y Survey research conclusions about? Field research Who will be observed Content analysis Existing data research for that purpose? Comparative research How will we actually Evaluation research measure the variables under study? Transforming the data collected into a form appropriate to manipulation and analysis FIGURE 4-5 The Research Process. Here are some of the key elements that we'll be examining throughout this book: the pieces that make up the whole of social research.

How to Design aResearch Project 109 provides a context for the specific details of particu- So, where do you start? You have an idea you lar components of social research. Essentially, it is want to pursue, one that involves abortion attitudes another and more detailed picture of the scientific and choice of college majoc In terms of the options process presented in Chapter 2. we've discussed in this chapter, you probably have both descriptive and explanatory interests, but you At the top of the diagram are interests, ideas, might decide you only want to explore the issue. and theories, the possible beginning points for a You might wonder what sorts of attitudes students line of research. The letters (A, B, X, Y, and so with different majors have about abortion (explo- forth) represent variables or concepts such as prej- ratory), what percentage of the student body sup- udice or alienation. Thus, you might have a general ports a woman's right to an abortion (descriptive), interest in finding out what causes some people to or what causes some to support it and others to be more prejudiced than others, or you might want oppose it (explanation). The units of analysis in this to know some of the consequences of alienation. case would individuals: college students. But we're Alternatively, your inquiry might begin with a jumping the gun. As you can see, even before specific idea about the way things are. For example, we've \"started,\" we've started. The reciprocal pro- you might have the idea that working on an assem- cesses described in Figure 4-5 begin even before bly line causes alienation. The question marks in you've made a commitment to a project. Let's look the diagram indicate that you aren't sure things are more formally at the various steps, then, keeping the way you suspect they are-that's why you're this reciprocal motion in mind. doing the research. Notice that a theory is repre- sented as a set of complex relationships among Getting Started several variables. At the outset of your project, then, your aim would The double arrows between \"interest,\" \"idea,\" probably be eX1Jloratory. At this point, you might and \"theory\" suggest that there is often a move- choose among several possible activities in pursu- ment back and forth across these several possible ing your interest in student attitudes about abor- beginnings. An initial interest may lead to the for- tion rights. To begin with, you might want to read mulation of an idea, which may be fit into a larger something about the issue. If you have a hunch theory, and the theory may produce new ideas and that attitudes are somehow related to college ma- create new interests. jor, you might find out what other researchers may have written about that. Appendix A of this book Any or all of these three may suggest the need will help you make use of your college library. In for empirical research. The purpose of such re- addition, you would probably talk to some people search can be to explore an interest, test a specific who support abortion rights and some who don't. idea, or validate a complex theory. Whatever the You might attend meetings of abortion-related purpose, the researcher needs to make a variety groups. All these activities could help prepare you of decisions, as indicated in the remainder of the to handle the various decisions of research design diagram. we're about to examine. To make this discussion more concrete, let's Before designing your study, you must define take a specific research example. Suppose you're the purpose of your project. What kind of study concerned with the issue of abortion and have a will you undertake-exploratory, descriptive, ex- special interest in learning why some college stu- planatory? Do you plan to write a research paper to dents support abortion rights and others oppose satisfy a course or thesis requirement? Is your pur- them. Going a step further, let's say you've formed pose to gain information that will support you in the impression that students in the humanities and arguing for or against abortion rights? Do you want social sciences seem generally more inclined to to write an article for the campus newspaper or an support the idea of abortion rights than those in the natural sciences do. (That kind of thinking often leads people to design and conduct social research.)

11 0 Chapter 4: Research Design academic journal? In reviewing the previous re- Similarly. you'll need to specify exact meanings search literature regarding abortion rights, you for all the other concepts you plan to studyo If you should note the design decisions other researchers want to study the relationship of opinion about have made, always asking whether the same deci- abortion to college major, you'll have to decide sions would satisfy your purposeo whether you want to consider only officially de- clared majors or to include students' intentions Usually, your purpose for undertaking research as welL What will you do with those who have can be expressed as a report. A good first step in no major? designing your project is to outline such a report (see Chapter 17 for help on this). Although your In surveys and experiments, you need to final report may not look much like your initial specify such concepts in advanceo In less tightly image of it this exercise will help you figure out structured research, such as open-ended inter- which research designs are most appropriate. Dur- views, an important part of the research may in- ing this step, clearly describe the kinds of state- volve the discovery of different dimensions, aspects, ments you want to make when the research is or nuances of conceptso In such cases, the research completeo Here are some examples of such state- itself may uncover and report aspects of social ments: \"Students frequently mentioned abortion life that were not evident at the outset of the rights in the context of discussing social issues that projecL concerned them personally.\" \"X percent of State U students favor a woman's right to choose an abor- Choice of Research Method tion.\" \"Engineers are (morelless) likely than sociol- ogists to favor abortion rights.\" As we'll discuss in Part 3, each research method has its strengths and weaknesses, and certain Conceptualization concepts are more appropriately studied through some methods than through otherso In our study Once you have a well-defined purpose and a clear of attitudes toward abortion rights, a survey might description of the kinds of outcomes you want to be the most appropriate method: either interview- achieve, you can proceed to the next step in the de- ing students or asking them to fill out a question- sign of your study-conceptualizationo We often naire. Surveys are particularly well suited to the talk pretty casually about social science concepts study of public opiniono This is not to say that you such as prejudice, alienation, religiosity, and liberal- couldn't make good use of the other methods ism, but it's necessary to clarify what we mean by presented in Part 3. For example, you might use these concepts, in order to draw meaningful con- the method of content analysis to examine letters clusions about themo Chapter 5 examines this pro- to the editor and analyze the different images of cess of conceptualization in deptho For now, let's see abortion that letter writers have. Field research what it might involve in the case of our hypotheti- would provide an avenue to understanding how cal exampleo people interact with one another regarding the is- sue of abortion, how they discuss it, and how If you're going to study how college students they change their mindso Other research methods feel about abortion and why, the first thing you'll introduced in Part 3 could also be used in studying have to specify is what you mean by \"the right to this topic. Usually, the best study design uses more an abortion.\" Because support for abortion proba- than one research method, taking advantage of bly varies according to the circumstances, you'll their different strengthso If you look back at the want to pay attention to the different conditions brief examples of actual studies at the end of the under which people might approve or disapprove preceding section, you'll see several instances of abortion: for example, when the woman's life is where the researchers used many methods in a in danger. in the case of rape or incest or simply as single study. a matter of personal choiceo

How to Design a Research Project 111 Operationalization decisions about population and sampling are re- lated to decisions about the research method to be Once you've specified the concepts to be studied usedo Whereas probability sampling techniques and chosen a research method, the next step is op- would be relevant to a large-scale surveyor a erationalization, or deciding on your measurement content analysis, a field researcher might need to techniques (discussed further in Chapters 5 and 6). select only those informants who will yield a bal- The meaning of variables in a study is determined anced picture of the situation under study, and in part by how they are measured, Part of the task an experimenter might assign subjects to experi- here is deciding how the desired data will be col- mental and control groups in a manner that creates lected: direct observation, review of official docu- comparability. ments, a questionnaire, or some other technique. In your hypothetical study of abortion atti- If you decided to use a survey to study attitudes tudes, the relevant population would be the stu- toward abortion rights, part of operationalization is dent population of your collegeo As you'll discover determining the wording of questionnaire itemso in Chapter 7, however, selecting a sample will re- For example, you might operationalize your main quire you to get more specific than thaL Will you variable by asking respondents whether they include part-time as well as full-time students? would approve of a woman's right to have an abor- Only degree candidates or everyone? International tion under each of the conditions you've conceptu- students as well as US. citizens? Undergraduates, alized: in the case of rape or incest, if her life were graduate students, or both? There are many such threatened by the pregnancy, and so forth. You'd questions-each of which must be answered in design the questionnaire so that it asked respon- terms of your research purposeo If your purpose is dents to express approval or disapproval for each to predict how students would vote in a local refer- situationo Similarly, you would specify exactly how endum on abortion, you might want to limit your respondents would indicate their college major, as population to those eligible and likely to voteo well as what choices to provide those who have not declared a major. Observations Population and Sampling Having decided what to study among whom by what method, you're now ready to make observa- In addition to refining concepts and measurements, tions-to collect empirical datao The chapters of you must decide whom or what to study The popu- Part 3, which describe the various research meth- lation for a study is that group (usually of people) ods, give the different observation techniques ap- about whom we want to draw conclusions. We're propriate to each. almost never able to study all the members of the population that interests us, however, and we can To conduct a survey on abortion, you might never make every possible observation of them. In want to print questionnaires and mail them to a every case, then, we select a sample from among sample selected from the student body. Alterna- the data that might be collected and studiedo The tively, you could arrange to have a team of inter- sampling of information, of course, occurs in every- viewers conduct the survey over the telephone. day life and often produces biased observations. The relative advantages and disadvantages of these (Recall the discussion of \"selective observation\" in and other possibilities are discussed in Chapter 9. Chapter L) Social researchers are more deliberate in their sampling of what will be observedo Data Processing Chapter 7 describes methods for selecting Depending on the research meth~d chosen, you'll samples that adequately reflect the whole popula- have amassed a volume of observations in a form tion that interests us. Notice in Figure 4-5 that that probably isn't immediately interpretable. If

112 Chapter 4: Research Design you've spent a month observing a street-corner Application gang firsthand, you'll now have enough field notes to fill a book. In a historical study of ethnic diver- The final stage of the research process involves the sity at your school, you may have amassed volumes uses made of the research you've conducted and of official documents, interviews with administra- the conclusions you've reached. To start, you'll tors and others, and so forth. Chapters 13 and 14 probably want to communicate your findings so describe some of the ways social scientific data are that others will know what you've learned. It may processed or transformed for qualitative or quanti- be appropriate to prepare-and even publish-a tative analysis. written report. Perhaps you'll make oral presenta- tions, such as papers delivered to professional In the case of a survey, the \"raw\" observations and scientific meetings. Other students would also are typically in the form of questionnaires with be interested in hearing what you've learned boxes checked, answers written in spaces, and the about them. like. The data-processing phase of a survey typically involves the classification (coding) of written-in You may want to go beyond simply reporting answers and the transfer of all information to a what you've learned to discussing the implications computer. of your findings . Do they say anything about ac- tions that might be taken in support of policy Analysis goals? Both the proponents and the opponents of abortion rights would be interested. Once the collected data are in a suitable form, you're ready to interpret them for the purpose of Finally, be sure to consider what your research drawing conclusions that reflect the interests, ideas, suggests in regard to further research on your sub- and theories that initiated the inquiry. Chapters 13 ject. What mistakes should be corrected in future and 14 describe a few of the many options available studies'? What avenues-opened up slightly in to you in analyzing data. In Figure 4-5, notice that your study-should be pursued further? the results of your analyses feed back into your initial interests, ideas, and theories. Often this Research Design in Review feedback represents the beginning of another cycle of inquiry. As this overview shows, research design involves a set of decisions regarding what topic is to be studied In the survey of student attitudes about abor- among what population with what research meth- tion rights, the analysis phase would pursue both ods for what purpose. Although you'll want to con- descriptive and explanatory aims. You might begin sider many ways of studying a subject-and use by calculating the percentages of students who fa- your imagination as well as your knowledge of a vored or opposed each of the several different ver- variety of methods-research design is the process sions of abortion rights. Taken together, these sev- of focusing your perspective for the purposes of a eral percentages would provide a good picture of particular study. student opinion on the issue. If you're doing a research project for one of Moving beyond simple description, you might your courses, many aspects of research design may describe the opinions of subsets of the student be specified for you in advance, induding the body, such as different college majors. Provided method (such as an experiment) or the topic (as that your design called for trapping other informa- in a course on a particular subject, such as preju- tion about respondents, you could also look at dice). The following summary assumes that you're men versus women; freshmen, sophomores, jun- free to choose both your topic and your research iors, seniors, and graduate students; or other cate- strategy. gories that you've included. The description of subgroups could then lead you into an explanatory In designing a research project, you'll find it analysis. useful to begin by assessing three things: your in- terests, your abilities, and the available resources.

The Research Proposal 113 Each of these considerations will suggest a large Once you have a general idea of what you number of possible studies. want to study and how, carefully review previous research in journals and books to see how other re- Simulate the beginning of a somewhat conven- searchers have addressed the topic and what they tional research project: Ask yourself what you're have learned about it. Your review of the literature interested in understanding. Surely you have sev- may lead you to revise your research design: Per- eral questions about social behavior and attitudes. haps you'll decide to use a previous researcher'S Why are some people politically liberal and others method or even replicate an earlier study. A stan- politically conservative? Why are some people dard procedure in the physical sciences, the inde- more religious than others? Why do people join pendent replication of research projects is just as militia groups? Do colleges and universities still dis- important in the social sciences, although social re- criminate against minority faculty members? Why searchers tend to overlook that. Or, you might would a woman stay in an abusive relationship? want to go beyond replication and study some as- Spend some time thinking about the kinds of ques- pect of the topic that you feel previous researchers tions that interest and concern you. have overlooked. Once you have a few questions you'd be inter- Here's another approach you might take. Sup- ested in answering for yourself, think about the pose a topic has been studied previously using field kind of information needed to answer them. What research methods. Can you design an experiment research units of analysis would provide the most that would test the findings those earlier re- relevant information: college students, corpora- searchers produced? Or, can you think of existing tions, voters, cities, or corporations? This question statistics that could be used to test their conclu- vvill probably be inseparable in your thoughts from sions? Did a mass survey yield results that you'd the question of research topiCS. Then ask which as- like to explore in greater detail through on-the-spot pects of the units of analysis would provide the in- observations and in-depth interviews? The use formation you need in order to answer your re- of several different research methods to test the search question. same finding is sometimes called triangulation, and you should always keep it in mind as a valuable Once you have some ideas about the kind of research strategy. Because each research method information relevant to your purpose, ask yourself has particular strengths and weaknesses, there is how you might go about getting that information. always a danger that research findings will reflect, Are the relevant data likely to be already available at least in part, the method of inquiry. In the best somewhere (say, in a government publication), or of all worlds, your own research design should would you have to collect them yourself? If you bring more than one research method to bear on think you would have to collect them, how would the topic. you go about doing it? Would you need to survey a large number of people or interview a few people The Research Proposal in depth? Could you learn what you need to know by attending meetings of certain groups? Quite often, in the design of a research project, Could you glean the data you need from books in you'll have to layout the details of your plan for the library? someone else's review and/or appr~val. In the case of a course project, for example, your instructor As you answer these questions, you'll find might very well want to see a \"proposal\" before yourself well into the process of research design. you set off to work. Later in your career, if you Keep in mind your own research abilities and the wanted to undertake a major project, you might resources available to you. There's little point in need to obtain funding from a foundation or gov- designing a perfect study that you can't actually ernment agency, who would most definitely want a carry out. You may want to try a research method you haven't used before so you can learn from it, but be careful not to put yourself at too great a disadvantage.

114 Chapter 4: Research Design detailed proposal that desCTibes how you would At the same time, you should focus your search on spend their money. You may respond to a Request your research topic: regardless of the methods for Proposals (RFP), which both public and private other researchers have used. So, if you're planning agencies often circulate in search of someone to do field research on, say, interracial marriages, you research for them. might gain some useful insights from the findings of surveys on the topic; further, past field research This chapter concludes 'with a brief discussion on interracial marriages could be invaluable in of how you might prepare a research proposaL This your designing a survey on the topic will give you one more overview of the whole re- search process that the rest of this book details. Because the literature review will appear early in your research proposal. you should write it vvith Elements ofaResearch Proposal an eye to introducing the reader to the topic you will address, laying out in a logical manner what Although some funding agencies (or your instruc- has already been learned on the topic by past re- tor, for that matter) may have specific requirements searchers, then leading up to the holes or loose ends for the elements or structure of a research proposal. in our knowledge of the topic, which you propose here are some basic elements you should include. to remedy. Or a little differently, your review of the literature may point to inconsistencies or disagree- Problem or Objective ments to be found among the existing research findings. In that case, your proposed research will What exactly do you want to study? Why is it aim to resolve the ambiguities that plague us. I don't worth studying? Does the proposed study have know about you, but I'm already excited about the practical significance? Does it contribute to the research you're proposing to undertake. construction of social theories? Subjects for Study Literature Review Whom or what will you study in order to collect What have others said about this topic? What theo- data? Identify the subjects in general. theoretical ries address it and what do they say? What previ- terms; in specific. more concrete terms, identify ous research exists? Are there consistent findings, who is available for study and how you'll reach or do past studies disagree? Are there flaws in the them. Will it be appropriate to select a sample? If body of existing research that you think you can so, how will you do that? If there is any possibility remedy? that your research will affect those you study, how will you insure that the research does not Chapter 17 has a lengthier discussion of this harm them? topic You'll find that special skills involved in read- ing social science research reports requires special Beyond these general questions, the specific re- skills. If you need to undertake a review of the lit- search method you'll use will further specify the erature at this point in your course, you may want matter. If you're planning to undertake an experi- to skip ahead to Chapter 17. It will familiarize you ment, a survey, or field research, for example, the v'lith the different types of research literature, how techniques for subject selection will vary quite a to find what you want, and how to read it. There is biL Happily, Chapter 7 of this book discusses sam- a special discussion of how to use electronic re- pling techniques for both qualitative and quantita- sources online and how to avoid being misled by tive studies. information on the Internet. Measurement In part, your review of the literature 'will be shaped by the data-collection method(s) you intend What are the key variables in your study? How will to use in your study. Reviewing the designs of pre- you define and measure them? Do your definitions vious studies using that same technique can give and measurement methods duplicate or differ from you a head start in planning your own study.

Main Points 115 those of previous research on this topic? If you As you can see, if you're interested in conduct- have already developed your measurement device ing a social research project. it's a good idea to pre- (a questionnaire, for example) or will be using pare a research proposal for your own purposes, something previously developed by others, it might even if you aren't required to do so by your in- be appropriate to include a copy in an appendix to structor or a funding agency If you're going to in- your proposaL vest your time and energy in such a project. you should do what you can to insure a return on that Data-Collection Methods investment. How vvill you actually collect the data for your Now that you've had a broad overview of social study? Will you conduct an experiment or a sur- research, let's move on to the remaining chapters in vey? Will you undertake field research or will you this book and learn exactly how to design and exe- focus on the reanalysis of statistics already created cute each specific step. If you've found a research by others? Perhaps you'll use more than one topic that really interests you, you'll want to keep it method. in mind as you see how you might go about study- ing it. Analysis MAIN POINTS Indicate the kind of analysis you plan to conduct. Introduction Spell out the purpose and logic of your analysis. I> Any research design requires researchers to Are you interested in precise description? Do you intend to explain why things are the way they are? specify as clearly as possible what they want Do you plan to account for variations in some qual- to find out and then determine the best way ity: for example, why some students are more lib- to do it. eral than others? What possible explanatory vari- ables will your analysis consider, and how vviIl you Three Purposes of Research know if you've explained variations adequately? I> The principal purposes of social research in- Schedule clude exploration, description, and explanation. Research studies often combine more than one It's often appropriate to provide a schedule for the purpose. various stages of research. Even if you don't do this I> Exploration is the attempt to develop an initial. for the proposal. do it for yourself. Unless you have rough understanding of some phenomenon. a timeline for accomplishing the several stages of I> Description is the precise measurement and re- research and keeping track of how you're doing, porting of the characteristics of some popula- you may end up in trouble. tion or phenomenon under study. I> Explanation is the discovery and reporting Budget of relationships among different aspects of the phenomenon under study. Whereas de- When you ask someone to cover the costs of your scriptive studies answer the question \"What's research, you need to provide a budget that so?\" ex-planatory ones tend to answer the specifies where the money vvill go. Large, expen- question \"Why?\" sive projects include budgetary categories such as personnel, equipment, supplies, telephones, and The Logic of Nomothetic Explanation postage. Even for a project you'll pay for yourself. I> Both idiographic and nomothetic models of it's a good idea to spend some time anticipating ex- penses: office supplies, photocopying, CD-ROMs, explanation rest on the idea of causation. telephone calls, transportation, and so on.

116 Chapter 4: Research Design The idiographic model aims at a complete un- @ Cross-sectional studies are based on observa- derstanding of a particular phenomenon, using tions made at one tin1eo Although such studies all relevant causal factorso The nomothetic are limited by this characteristic, researchers model aims at a general understanding-not can sometimes make inferences about pro- necessarily complete-of a class of phenom- cesses that occur over timeo ena, using a small number of relevant causal factorso @ In longitudinal studies, observations are made at @ There are three basic criteria for establishing many tin1es. Such observations may be made of causation in nomothetic analyses: (1) The vari- samples drawn from general populations (trend ables must be empirically associated, or corre- studies), samples drawn from more specific sub- lated, (2) the causal variable must occur earlier populations (cohort studies), or the same in time than the variable it is said to affect, and sample of people each time (panel studies)o (3) the observed effect cannot be explained as the effect of a different variable. How to Design a Research Project @ Research design starts with an initial interest, Necessary and Sufficient Causes @ Mere association, or correlation, does not in it- idea, or theoretical ex-pectation and proceeds through a series of interrelated steps to narrow self establish causationo A spurious causal rela- the focus of the study so that concepts, meth- tionship is an association that in reality is ods, and procedures are well definedo A good caused by one or more other variables. research plan accounts for all these steps in advance. Units of Analysis @ At the outset, a researcher specifies the mean- @ Units of analysis are the people or things whose ing of the concepts or variables to be studied (conceptualization), chooses a research method characteristics social researchers observe, de- or methods (e.go, experiments versus surveys), scribe, and explaino Typically, the unit of analy- and specifies the population to be studied and, sis in social research is the individual person, if applicable, how it ,viII be sampledo but it may also be a social group, a formal orga- nization, a social interaction, a social artifact, or @ To operationalize the concepts to be studied, some other phenomenon such as a lifestyle or a the researcher states precisely how variables in type of social interaction. the study will be measuredo Research then pro- @ The ecological fallacy involves conclusions ceeds through observation, data processing, dravvn from the analysis of groups (e.go, corpo- analysis, and application, such as reporting the rations) that are then assumed to apply to indi- results and assessing their implicationso viduals (e.g., the employees of corporations). @ Reductionism is the attempt to understand a The Research Proposal complex phenomenon in terms of a narrow set @ A research proposal provides a preview of why of concepts, such as attempting to explain the American Revolution solely in terms of eco- a study will be undertaken and how it vvill be nomics (or political idealism or psychology). conducted. A research project is often required to get permission or necessary resourceso Even when not required, a proposal is a useful de- vice for planning. The Time Dimension KEY TERMS @ Research into processes that occur over time The following terms are defined in context in the presents social challenges that can be addressed chapter and at the bottom of the page where the term through cross-sectional studies or longitudinal studies.

Additional Reading 117 is introduced, as well as in the comprehensive glossary L The analysis of community opposition at the back of the booko to group homes for the mentally handi- capped 00. indicates that deteriorating cohort study reductionism neighborhoods are most likely to organize correlation social artifact in opposition, but that upper-middle class cross-sectional study sociobiology neighborhoods are most likely to enjoy ecological fallacy spurious relationship private access to local officialso (Graham longitudinal study trend study and Hogan 1990: 513) panel study units of analysis go Some analysts during the 1960s predicted REVIEW QUESTIONS AND EXERCISES that the rise of economic ambition and po- litical militancy among blacks would foster 10 One example in this chapter suggested that po- discontent with the \"otherworldly\" black litical orientations cause attitudes toward legal- mainline churcheso (Ellison and Sherkat izing marijuana. Can you make an argument 1990: 551) that the time order is just the opposite of what was assumed? IL This analysis explores whether propositions and empirical findings of contemporary 2. Here are some examples of real research theories of organizations directly apply to topics. For each one, can you name the unit of both private product producing organiza- analysis? (The answers are at the end of this tions (PPOs) and public human service chapteL) organizations (PSOs)o (Schiflett and ley 1990: 569) a. Women watch TV more than men because they are likely to work fewer hours outside L This paper examines variations in job title the home than meno 00 Black people watch structures across work roleso Analyzing an average of approximately three-quarters 3,173 job titles in the California civil service of an hour more television per day than system in 1985, we investigate how and white people. (Hughes 1980: 290) why lines of work vary in the proliferation of job categories that differentiate ranks, b. Of the 130 incorporated UoSo cities with functions, or particular organizationalloca- more than 100,000 inhabitants in 1960, tionso (Strang and Baron 1990: 479) 126 had at least two short-term nonpropri- etary general hospitals accredited by the 3 Review the logic of spuriousnesso Can you think American Hospital Associationo (Turk up an example where an observed relationship 1980: 317) between two variables could actually be ex- plained away by a third variable? c The early TM [transcendental meditation] organizations were small and informal. The 40 Using InfoTrac College Edition or printed jour- Los Angeles group, begun in June 1959, nals in the library, locate a research project in- met at a member's house where, inciden- volving a panel studyo Describe the nature of the tally, Maharishi was living. (Johnston study design and its primary findingso 1980: 337) ADDITIONAL READINGS d.. However, it appears that the nursing staffs exercise strong influence over 000a decision Bart, Pauline, and Linda FrankeL 19860 The Student to change the nursing care systemo . 00 Sociologist's Handbook.. Morristown, NJ: General Conversely, among those decisions domi- Learning Press. A handy little reference book to nated by the administration and the medical help you get started on a research projecL Writ- staffs 000(Comstock 1980: 77) ten from the standpoint of a student term pa- per, this volume offers a particularly good guide e.. Though 667,000 out of 2 million farmers in to the periodical literature of the social sciences the United States are women, women his- that's available in a good libraryo torically have not been viewed as farmers, but rather, as the farmer's wifeo (Votaw Casley, D. J., and Do Ao Luryo 19870 Data Collection in 1979: 8) Developing COlllltrieso Oxford: Clarendon Presso

118 Chapter 4: Research Design This book discusses the special problems of re- Online Study Resources search in the developing world. Sociology @Now'\": Research Methods Cooper, Harris M. 1989. Imegrating Research A Guide for Literature RevielVs. Newbury Park, CA: Sage. 1. Before you do your final review of the chapter, The author leads you through each step in the take the SociologyNolF Researclz Merlzods diagnos- literature review process. tic quiz to help identify the areas on which you should concentrate. You'll find information on Hunt, Morron. 1985. Profiles ofSocial Research The this online tooL as well as instructions on how SciClltific SlIIdy of Human Imeracrions . New York: to access all of its great resources, in the front of Basic Books. An engaging and informative the book. series of project biographies: James Coleman'S study of segregated schools is presented, as 2. As you review, take advantage of the Sociology well as several other major projects that illus- Now Research lYlerhods customized study plan, trate the elements of social research in actual based on your quiz results. Use this study plan practice. with its interactive exercises and other re- sources to master the materiaL Iversen, Gudmund R. 1991. ComexlIIal Analysis. Newbury Park, CA: Sage. Contextual analysis 3. When you're finished with your review, take examines the impact of socioenvironmental fac- the posttest to confirm that you're ready to tors on individual behavior. Durkheim's study of move on to the next chapter. suicide offers a good example of this, identifying social contexts that affect the likelihood of self- WEBSITE FOR THE PRACTICE destruction. OF SOCIAL RESEARCH 11 TH EDITION Maxwell, Joseph A. 1996. Qualirative Research Go to your book's website at http://sociology Design: An Interacriw Approaciz. Newbury Park, .wadsworth.com/babbie_practicelle for tools to CA: Sage. Maxwell covers many of the same aid you in studying for your exams. You'll find Tilto- topics that this chapter does but with attention rial Quizzes with feedback, Imenzet Ewrcises, Flaslzcards, devoted specifically to qualitative research and Chapter TlItorials, as well as E\\1ended Projects, Info- projects. Trae College Edition search terms, Social Research in Cyber- space, GSS Data, H'c?b Links, and primers for using vari- Menard, Scott. 1991 . Longillidinal Research. Newbury ous data-analysis software such as SPSS and l',TVivo. Park, CA: Sage. Beginning by explaining why researchers conduct longitudinal research, the WEB LINKS FOR THIS CHAPTER author goes on to detail a variety of study de- signs as well as suggestions for the analysis of Please realize that the Internet is an evolv- longitudinal data ing entity, subject to change. Nevertheless, these few websites should be fairly stable. Miller, Delbert. 1991. Handbook of Researciz Design Also, check your book's website for even more H't?b and Social MeasuremCllf. Newbury Park. CA: Links Sage. A useful reference book for introducing or reviewing numerous issues involved in de- The Internet Public Library, Social Sciences sign and measurement. In addition, the book Resources contains a wealth of practical information relat- http://www,ipLorg/reflRR/static/socOO.OO.OO.html ing to foundations, journals, and professional This site, along with its numerous hotlinks, provides a associations. broad view of the kinds of research topics explored by social researchers in many disciplines. SPSS EXERCISES University of Calgary, Beginner's Guide See the booklet that accompanies your text for to the Research Proposal exercises using SPSS (Statistical Package for the http://www.ucalgary.ca/md/CAH/research/ Social Sciences). There are exercises offered for each res_prop.htm chapter, and you'll also find a detailed primer on using SPSS.

As the name suggests, this site will walk you through Online Study Resources 119 the process of preparing a research proposaL c Transcendental meditation organizations Anthony W. Heath, \"The Proposal (groups) in Qualitative Research\" http://vvwwnova.edu/ssss/QR/QR3-l/heath.html d. Nursing staffs (groups) This piece, reprinted from The QlIalirative Report 3 (no. e, Farmers (individuals) 1, March 1997) provides another guide to proposal L Neighborhoods (groups) writing, this time specifically for qualitative research g. Blacks (individuals) projects. h. Service and production organizations (formal ANSWERS TO UNITS OF ANALYSIS QUIZ, EXERCISE 2 organizations) i. Job titles (artifacts) a.. Men and women, black and white people (individuals) b.. Incorporated US, cities (groups)

Conceptualization, Operationalization, and Measurement Introduction Variations between the Extremes Measuring Anything That Exists A Note on Dimensions Conceptions, Concepts, Defining Variables and Reality and Attributes Concepts as Constructs Levels of Measurement Single or Multiple Indicators Conceptualization Some Illustrations of Indicators and Dimensions Operationalization Choices The Interchangeability Operationalization Goes of Indicators On and On Real, Nominal, and Operational Definitions Criteria of Measurement Quality Creating Conceptual Order Precision and Accuracy Reliability An Example Validity of Conceptualization: Who Decides What's Valid? The Concept of Anomie Tension between Reliability and Validity Definitions in Descriptive and Explanatory Studies Operationalization Choices Range of Variation SociologyWNow'\": Research Methods Use this online tool to help you make the grade on your next exam. After reading this chapter, go to the \"Online Study Resources\" at the end of the chapter for instructions on how to benefit from SodologyNow: Research Methods\"

Measuring Anything That Exists 121 Introduction you had found yourself being misclassified by the measurement system. This chapter and the next deal with how re- searchers move from a general idea about what Your feeling of dissatisfaction reflects an impor- they want to study to effective and well-defined tant fact about social research: Most of the variables measurements in the real world. This chapter dis- we want to study don't actually exist in the way cusses the interrelated processes of conceptualiza- that rocks exist. Indeed, they are made up. More- tion, operationalization, and measurement. Chapter over, they seldom have a single, unambiguous 6 builds on this foundation to discuss types of meaning. measurements that are more complex\" To see what I mean, suppose we want to study We begin this chapter by confronting the hidden political parry affiliatiolL To measure this variable, we concern people sometimes have about whether it's might consult the list of registered voters to note truly possible to measure the stuff of life: love, hate, whether the people we were studying were regis- prejudice, religiosity, radicalism, alienation. The tered as Democrats or Republicans and take that as answer is yes, but it will take a few pages to see how. a measure of their party affiliation. But we could Once we establish that researchers can measure also simply ask someone what party they identify anything that exists, we'll turn to the steps involved with and take their response as our measure. No- in doing just that. tice that these two different measurement possibili- ties reflect somewhat different definitions of \"politi- Measuring Anything cal party affiliation.\" They might even produce That Exists different results: Someone may have registered as a Democrat years ago but gravitated more and more Earlier in this book, I said that one of the two pillars toward a Republican philosophy over time\" Or of science is observation\" Because this word can sug- someone who is registered with neither political gest a casuaL passive activity, scientists often use the party may, when asked, say she is affiliated with term measurement instead, meaning carefuL deliber- the one she feels the most kinship with. ate observations of the real world for the purpose of describing objects and events in terms of the attri- Similar points apply to religiolls affiliatioll. Some- butes composing a variable. times this variable refers to official membership in a particular church, temple, mosque, and so forth; You may have some reservations about the abil- other times it simply means whatever religion, if ity of science to measure the really important as- any, you identify yourself with. Perhaps to you it pects of human social existence. If you've read means something else, such as attendance at research reports dealing vvith something like liber- religious services. alism or religion or prejudice, you may have been dissatisfied with the way the researchers measured The truth is that neither \"party affiliation\" nor whatever they Viere studying. You may have felt \"religious affiliation\" has any real meaning, if by that they were too superficiaL that they missed the \"real\" we mean corresponding to some objective aspects that really matter most. Maybe they mea- aspect of reality. These variables do not exist in na- sured religiosity as the number of times a person ture. They are merely terms we've made up and as- went to religious services, or maybe they measured signed specific meanings to for some purpose, such liberalism by how people voted in a single election. as doing social research. Your dissatisfaction would surely have increased if But, you might object, \"political affiliation\" and \"religious affiliation\"-and a host of other things social researchers are interested in, such as prej- udice or compassion-have some reality. After all, we make statements about them, such as \"In

122 Chapter 5: Conceptualization, Operationalization, and Measurement Happytown, 55 percent of the adults affiliate with observed it We just agreed to use it as a shortcut, a the Republican Party, and 45 percent of them are name that represents a collection of apparently re- Episcopalians. Overall, people in Happytown are lated phenomena that we've each observed in the low in prejudice and high in compassion.\" Even course of life. In short, we made it up. ordinary people, not just social researchers, have been known to make statements like that If these Here's another clue that prejudice isn't some- things don't exist in reality, what is it that we're thing that exists apart from our rough agreement measuring and talking about? to use the term in a certain way.. Each of us devel- ops our own mental image of what the set of real What indeed? Let's take a closer look by phenomena we've observed represents in general considering a variable of interest to many social and what these phenomena have in common. researchers (and many other people as well)- When I say the word prejudice, it evokes a mental prejudice. image in your mind, just as it evokes one in mine. It's as though file drawers in our minds contained Conceptions, Concepts, and Reality thousands of sheets of paper, with each sheet of pa- per labeled in the upper right-hand corner. A sheet As you and I wandered dovvn the road of life, we of paper in each of our minds has the term prejudice observed a lot of things and knew they were real on it. On your sheet are all the things you've been through our observations, and we heard reports told about prejudice and everything you've ob- from other people that seemed real. For example: served that seems to be an example of it. My sheet has what I've been told about it plus all the things e We personally heard people say nasty things I've observed that seem examples of it-and mine about minority groups. isn't the same as yours. e We heard people say that women were inferior The technical term for those mental images, to men. those sheets of paper in our mental file drawers, is concepcion. That is, I have a conception of prejudice, e We read about African Americans being and so do you. VVe can't communicate these mental lynched. images directly, so we use the terms written in the upper right-hand corner of our own mental sheets e We read that women and minorities earned less of paper as a way of conm1Unicating about our con- for the same work. ceptions and the things we observe that are related to those conceptions. These terms make it possible e We learned about \"ethnic cleansing\" and wars for us to communicate and eventually agree on in which one ethnic group tried to eradicate what we specifically mean by those terms. In social another. research, the process of coming to an agreement about what terms mean is conceptualization, and the With additional experience, we noticed some- result is called a concept. thing more. People who participated in lynching were also quite likely to call African Americans Let's take another example of a conception. ugly names. A lot of them, moreover, seemed to Suppose that I'm going to meet someone named want women to \"stay in their place.\" Eventually it Pat, whom you already know. I ask you what Pat is dawned on us that these several tendencies often like. Now suppose that you've seen Pat help lost appeared together in the same people and also had children find their parents and put a tiny bird back something in common. At some point, someone in its nest. Pat got you to take turkeys to poor fami- had a bright idea: \"Let's use the word prejlldiced as a lies on Thanksgiving and to visit a children's hospi- shorthand notation for people like that. We can use talon Christmas. You've seen Pat weep through a the term even if they don't do all those things-as movie about a mother overcoming adversities to long as they're pretty much like that.\" save and protect her child. As you search through your mental files, you may find all or most of those Being basically agreeable and interested in efficiency, we agreed to go along with the system. That's where \"prejudice\" came from. We never

Measuring Anything That Exists 123 phenomena recorded on a single sheet labeled of us have different experiences and observations, \"compassionate.\" You look over the other entries no two people end up with exactly the same set of on the page, and you find they seem to provide entries on any sheet in their file systems. If we want an accurate description of Pat So you say, \"Pat is to measure \"prejudice\" or \"compassion,\" we must compassionate.\" first stipulate what, exactly, counts as prejudice or compassion for our purposes. Now I leaf through my own mental file drawer until I find a sheet marked \"compassionate.\" I then Returning to the assertion made at the outset of look over the things written on my sheet, and I say, this chapter, we can measure anything that's reaL \"Oh, that's nice.\" I now feel I know what Pat is like, We can measure, for example, whether Pat actually but my expectations reflect the entries on my file puts the little bird back in its nest, visits the hospital sheet, not yours. Later, when I meet Pat, I happen on Christmas, weeps at the movie, or refuses to con- to find that my own experiences correspond to the tribute to saving the whales. All of those behaviors entries I have on my \"compassionate\" file sheet, exist, so we can measure them. But is Pat really and I say that you sure were right. compassionate? We can't answer that question; we can't measure compassion in any objective sense, But suppose my observations of Pat contradict because compassion doesn't exist in the way that the things I have on my file sheet. I tell you that I those things I just described exist. Compassion ex- don't think Pat is very compassionate, and we begin ists only in the form of the agreements we have to compare notes. about how to use the term in communicating about things that are reaL You say, \"I once saw Pat weep through a movie about a mother overcoming adversity to save and Concepts as Constructs protect her child.\" I look at my \"compassionate sheet\" and can't find anything like that. Looking If you recall the discussions of postmodernism in elsewhere in my file, I locate that sort of phenome- Chapter 1, you'll recognize that some people would non on a sheet labeled \"sentimental.\" I retort, object to the degree of \"reality\" I've allowed in the \"That's not compassion. That's just sentimentality.\" preceding comments. Did Pat \"really\" visit the hos- pital on Christmas? Does the hospital \"really\" ex- To further strengthen my case, I tell you that I ist? Does Christmas? Though we aren't going to be saw Pat refuse to give money to an organization radically postmodern in this chapter, I think you'll dedicated to saving whales from extinction. \"That recognize the importance of an intellectually tough represents a lack of compassion,\" I argue. You view of what's real and what's not. (When the in- search through your files and find saving the tellectual going gets tough, the tough become social whales on two sheets-\"environmental activism\" scientists. ) and \"cross-species dating\"-and you say so. Even- tually, we set about comparing the entries we have In this context, Abraham Kaplan (1964) distin- on our respective sheets labeled \"compassionate.\" guishes three classes of things that scientists mea- We then discover that many of our mental images sure. The first class is direct observables: those things corresponding to that term differ. we can observe rather simply and directly, like the color of an apple or the check mark made in a In the big picture, language and communication questionnaire. The second class, indirect observables, work only to the extent that you and I have consid- require \"relatively more subtle, complex, or indi- erable overlap in the kinds of entries we have on our rect observations\" (1964: 55). We note a person's corresponding mental file sheets. The similarities we check mark beside \"female\" in a questionnaire and have on those sheets represent the agreements ex- have indirectly observed that person's gender. His- isting in our society. As we grow up, we're told ap- tory books or minutes of corporate board meetings proximately the same thing when we're first intro- provide indirect observations of past social actions. duced to a particular term. Dictionaries formalize Finally, the third class of observables consists of the agreements our society has about such terms. Each of us, then, shapes his or her mental images to correspond with such agreements. But because all

124 Chapter 5: Conceptualization, Operationalization, and Measurement construas-theoretical creations that are based on tured task of discovering what those real meanings observations but that cannot be observed directly are and what constitutes a genuine measurement of or indirectly. A good example is intelligence quo- them. Regarding constructs as real is called reifica- tient or IQ. It is constructed mathematically from [ion. The reification of concepts in day-to-day life is observations of the answers given to a large num- quite common. In science, we want to be quite clear ber of questions on an IQ test No one can directly about what it is we are actually measuring, but this or indirectly observe IQ. It is no more a \"real\" char- aim brings a pitfall with it. Settling on the \"best\" acteristic of people than is compassion or prejudice. way of measuring a variable in a particular study may imply that we've discovered the \"real\" meaning Kaplan (1964: 49) defines concept as a \"family of the concept involved. In fact concepts have no of conceptions.\" A concept is, as Kaplan notes, a reaL true, or objective meanings-only those we construct, something we create. Concepts such as agree are best for a particular purpose. compassion and prejudice are constructs created from your conception of them, my conception of Does this discussion imply that compassion, them, and the conceptions of all those who have prejudice, and similar constructs can't be mea- ever used these terms. They cannot be observed di- sured? Interestingly, the answer is no. (And a good rectly or indirectly, because they don't exist. We thing, too, or a lot of us social researcher types made them up. would be out of work.) I've said that we can mea- sure anything that's real. Constructs aren't real in To summarize, concepts are constructs derived the way that trees are reaL but they do have an- by mutual agreement from mental images (concep- other important virtue: They are useful. That is, tions) . Our conceptions sunm1arize collections of they help us organize, communicate about. and seemingly related observations and experiences. Al- understand things that are real. They help us make though the observations and experiences are reaL predictions about real things. Some of those predic- at least subjectively, conceptions, and the concepts tions even turn out to be true. Constructs can work derived from them, are only mental creations. The this way because, although not real or observable terms associated with concepts are merely devices in themselves, they have a definite relationship to created for the purposes of filing and communica- things that are real and observable. The bridge from tion. A term such as prejudice is, objectively speak- direct and indirect observables to useful constructs ing, only a collection of letters. It has no intrinsic is the process called conceptualization. reality beyond that Is has only the meaning we agree to give it Conceptualization Usually, however, we fall into the trap of be- As we've seen, day-to-day communication usually lieving that terms for constructs do have intrinsic occurs through a system of vague and general meaning, that they name real entities in the world. agreements about the use of terms. Although you That danger seems to grow stronger vvhen we be- and I do not agree completely about the use of the gin to take terms seriously and attempt to use them term compassionate, I'm probably safe in assuming precisely. Further, the danger is all the greater in that Pat won't pull the wings off flies. A wide range the presence of experts who appear to know more of misunderstandings and conflict-from the inter- than we do about what the terms really mean: It's personal to the international-is the price we pay easy to yield to authority in such a situation. for our imprecision, but somehow we muddle through. Science, however. aims at more than Once we assume that terms like prejudice and muddling; it cannot operate in a context of such compassion have real meanings, we begin the tor- imprecision. conceptualization The mental process whereby The process through which we specify what we fuzzy and imprecise notions (concepts) are made mean when we use particular terms in research is more specific and precise. So you want to study prej- called conceptualization. Suppose we want to udice. What do you mean by \"prejudice\"? Are there different kinds of prejudice? What are they?