• • • •• • •

• •• Is sociology scientific? The research process . - - ...... Causation and correlation

Research methods .. Ethnography

Experiments Biographical research Comparative research Historical analysis ComIJarative-historicaI researcli Sociological research in the real world Restating the obvious? The influence of sociology Sumllla ry {Joi Ilts Further reading Internet links

(opposite) Why hang out m pubJtc toilets? owards the end of a working day, the public toilets of a particular park in TSt Louis, Missouri, in the United States, are suddenly busier than one might expect. One man walks in dressed in a grey suit; another has on a base­ ball cap, trainers, shorts and a T-shirt; a third is wearing the mechanic's uniform from the garage where he has repaired cars all day. What are these men doing here? Surely other toilets are more conveniently located. Is there some common interest besides the toilets themselves that brings them to this place? None of these men is visiting the toilets to use them for the purposes for which they were officially built: they are there for 'instant sex'. Many ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS

the book caused widespread controversy when it was published, and for some people the issue is still a difficult one to deal with today. For example, Humphreys' research methodology was heavil y criticized for being unethical because his fieldwork had to be covert; it did not involve seeking the informed consent of the men he studied. However, his research in the 'tearooms' was able to cast new light on the struggles of men who were forced to keep their sexual proclivities secret. He showed that many men who otherwise live 'normal' li ves - the people next door - find ways to engage in embarrassing behaviours that will not harm their careers or family lives. Humphreys' research was conducted at a time when there was much more stigma associated with gay and lesbian identities and when police were vigil ant in enforcing laws In contrast to Humphrey's covert research, against such behaviour. Many lives were increasingly the subjects of research are ruined in the process of harsh law enforce­ themselves involved in the research process and may help to formulate questions or ment. Humphreys spent an extended period of comment on the researcher's interpretation of their views. time researching public toilets because one excell e nt way to understand social processes is to participate in and observe men - married and unmarried, those with them. He also conducted survey interviews straight identities and those who see them­ that enabled him to gather more informa­ selves as gay - seek sex with people they do tion than he would have obtained by simply not know. They are hoping to experience observing the toilets. Humphreys' research sexual excitement, but they want to avoid opened a window on an aspect of life that involvement. They do not want any many people were shocked to know existed commitments that extend beyond the at all and that certainly needed to be under­ particular encounters they will have in this stood at a deeper level. His work was based public convenience. on systematic research, but it also carried a This kind of search for anonymous, note of passion; it was conducted with a instant sex between men happens all over certain scientific detachment, but at the the world, yet until the late 1960s the same time Humphreys was involved in the phenomenon persisted as a widespread but search for solutions to social problems. rarely studied form of human interaction. In Humphreys argued that persecution of the USA, the gay community called the gay lifestyles leads men to live anguished toile ts where these activities occur existences in which they must resort to 'tearooms' (in the UK the activity is known extreme secrecy and often dangerous activ­ as 'cottaging'). Laud Humphreys, a sociolo­ ities. His study was conducted before the gist, visited these public lavatories to onset of AlDS; such activity would be much conduct research on the participants. He more dangerous today. He argued that toler­ wrote about them in his book Tearoom ance for a gay subculture would put gays in Trade (Humphreys 1970). Unsurprisingly, a position where they could provide one Asking and Answering Sociological Questions another with self-esteem, mutual support reveal the real reason for his presence in the and relief from torment. tearoom. Was this particular aspect of his behaviour ethical? The answer is that, on balance, this particular aspect of his study Human subjects and did not put any of his subjects at risk. On the basis of what he observed in the tearoom, ethical problems Humphreys did not collect information All research concerned with human beings about the participants that would have can pose ethical dilemmas. A key question identified them. What he knew about them that sociologists agree must be asked is was similar to what all the other people in whether the research 'poses risks to the the tearoom knew. In this way, his presence subjects that are greater than the risks they did not subject them to any more risk than face in their everyday lives'. Ethical issues they already encountered in their everyday have become much more prominent today lives. At the same time, had Humphreys than in the past. In particular, researchers been completely frank at every stage, the are no longer seen as the knowledgeable research might not have progressed as far as experts or participants as mere subjects of it did. Indeed, some of the most valuable research. Increasingly, the subjects of data that have been collected by sociologists research are themselves involved in the could never have been gathered if the research process and may help to formulate researcher had first explained the project to questions, comment on the researcher's each person encountered in the research interpretation of their views and in some process. cases expect to receive a copy of the final If this were the only dilemma posed by research report. Humphreys' research project, it would not Clearly, as is the case in other areas of stand out as a notable problem in the ethics social life (such as doctors and their patients of social research. What raised more or university professors and their students), eyebrows was that Humphreys wrote down 'lay people' no longer automatically defer to the car number-plates of the people who 'experts' in the way they might have done a came into the 'tearooms', obtained their few decades ago. This broad social change is home addresses from a friend who worked also transforming research practice. Indeed, at the Department of Motor Vehicles and all research funding bodies now routinely then visited their homes in the guise of ask research teams what ethical issues they conducting a neutral survey. Even though expect to confront and how they will deal Humphreys did not reveal to the men's fami­ with these, whether deception will be used, lies anything about their activities in the what measures will be put in place to tearooms, and even though he took great protect their participants from risk and how pains to keep the data confidential, the their findings will be fed back to partici­ knowledge he gained could have been pants at the end of the study. damaging. Since the activity he was docu­ In writing Tearoom Trade, Humphreys menting was illegal, police officers might said he was less than truthful to those whose have demanded that he release information behaviour he was studying. He said he did about the identities of the subjects. It is also not reveal his identity as a sociologist when possible that a less skilled investigator could observing the tearoom. People who came have slipped up when interviewing the into the tearoom assumed he was there for subjects' families, or that Humphreys could the same reasons they were and that his have lost his notes, which might then have presence could be accepted at face value. been found by someone else. Although he did not tell any direct lies Considering the number of things that during his observations, neither did he could go wrong in the research process, ASK ING AND ANSWERING SOCIOLOGICAL QUESTIONS researchers today do not consider projects obvious to all (see chapter 7, 'Social Interac­ of this kind to be legitimate. Around the tion and Everyday Life' and chapter 5, 'The world, government funding bodies for soci- Environment'). In this case, what most 010gica
research, such as the European people believed to be simply a public Science Foundation or the UK's Eco nomic building with an immediately obvious and Social Research Council (ESRC), as well function was, for a particular social group, as the professional organizations to which primarily a venue for the pursuit of sexual sociologists belong, such as the Interna­ satisfaction. tional Sociological Association (lSA) , now It is also interesting to note that elements have much stri cter ethical guidelines for of modern theoretical approaches can help researchers engaging in sociological us understand the issues addressed by research of whatever kind. Humphreys' study. An interactionist might Humphreys was one of the first sociolo­ ask: how does this behaviour take place gists to study the lives of gay men. His through processes of interaction? What account was a humane treatment that went kinds of interaction take place? Humphreys well beyond the existing stock of knowledge found that people who went into the on sexual communities. Nthough none of tearooms learnt from others to be silent. his research subjects actually suffered as a This was a response to the demand for result of his book, Humphreys himself later privacy without invo lvement. Another agreed with his critics on the key ethical finding was that men who went into the controversy. He said that were he to do the toilet and who did not respond to initial study again, he would not trace number­ sexual advances were not approached any plates or go to people's homes. Instead, after further. Each party must collaborate to gathering his data in the public 'tearooms', make a sexual situation occur. A functional­ he might try to get to know a subset of the ist approach might ask: what contribution people well enough to inform them of his does the tearoom make to the continuation goals for the study and then ask them to talk of society as a whole? The answer is that it about the significance of these activities in provides an outlet for sexual activity that, their lives. when carried out in secret, enables th e participants and other members of society to carry on as 'normal' people in their Sociological questions everyday lives without challenging the accepted order of things. A Marxist The toilets under study in Tearoom Trade approach might ask: is thinking about are a perfect example of a phenomenon that economic class relations apparent in the is the subject of many of the kinds of ques­ tearooms? Humphreys found that the tion that sociologists ask. For example, in impersonal sex of the tearooms had a looking at the surprising activities that democratic quality. Men of all social classes occur in public toilets, Humphreys was and races would come together in these asking how society works in ways that are places for sexual contact. Finally, a feminist different from the official versions of how it approach might ask: how can women's lives should work. He also found that what we be considered in this study of an all-male take to be naturai - a public toilet - is actu­ group? This approach was not dominant at ally socially constructed, depending on how the time Humphreys conducted his study, it is used. Social constructionism is a but a feminist today might ask how perspective which begins from the premise women's lives - perhaps wives and partners that social reality is - to varying degrees - who know nothing about the activity of the product of interactions between indi­ their mal e partners - are affected and put at viduals and groups, not something that is ri sk by the secret behaviour in the Law enforcement exists in all countries, but comparative empirical studies of police forces are needed to reveal their similarities and differences. tearooms. We explore some of these theo­ and in the findings it comes up with. The retical approaches in chapter 3. issues that concern sociologists, in both It is almost 40 years since Tearoom Trade their theorizing and their research, are often was first published and in the interim society similar to those that worry other people. But has become more tolerant of gay identities the results of such research frequently run and gay sex. After the publication of his book, counter to our common-sense beliefs. Humphreys became part of the political What are the circumstances in which movement - the gay rights movement - that racial or sexual minorities live? How can made this change possible. He used his mass starvation exist in a world that is far findings to convince courts and police to wealthier than it has ever been before?What ease up on prosecuting men for engaging in effects will the increasing use of the Internet gay sex so as to alleviate the damaging side have on our lives? [s the family beginning to effects of covert sexual activity. disintegrate as an institution? Sociologists It is the business of sociological research try to provide answers to these and many in general to go beyond surface-level under­ other questions. Their findings are by no standings of ordinary life, as Humphreys means conclusive. Neverth eless, it is always clearly did. Good research should help us the aim of sociological theorizing and understand our social lives better, some­ research to break away from the speculative times in a new way. It could take us by manner in which the ordinary person surprise, both in the questions that it asks usually considers such questions. Good ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS sociological work tries to make the ques­ cant differences, for example, between tions as precise as possible and seeks to the social and legal systems of the USA, gather factual evidence before coming to Italy and South Africa. A typical com­ conclusions. To achieve these aims, we must parative question might be: how much know the most useful research methods do patterns of criminal behaviour and to apply in a given study and how best to law enforcement vary between the two analyse the results. countries? In their research studies, sociologists will In sociology, we need not only to look at often ask empirical or factual questions. For existing societies in relation to one another example, many aspects of sexual behaviour. but also to compare their present and past. such as those Humphreys studied, need The questions sociologists ask in this case direct and systematic sociological investiga­ are historical or developmental questions. tion. Thus, we might ask: what kinds of How did we get from there to here? To occupation and domestic arrangement are understand the nature of the modern world, most common among people who go to the we have to look at previous forms of society tearooms? What proportion of tearoom and also study the main direction that participants do the police catch? Factual processes of change have taken. Thus we questions of this kind are often difficult to can investigate, for example. how the first answer. Official statistics on tearooms do prisons originated and what they are like not exist. Even official statistics on crime are today. of dubious value in revealing the real level of Factual investigations - or, as sociologists criminal activity. Researchers who have usually prefer to call them, empirical inves­ studied crime levels have found that only tigations - concern how things occur. Yet about half of all serious crimes are reported sociology does not consist of just collecting to the police. facts, however important and interesting Factual information about one society, they may be. We always need to interpret of course, will not always tell us whether we what facts mean, and to do so we must learn are dealing with an unusual case or a to pose theoretical questions - concerned general set of influences. Sociologists often with why things occur. Many sociologists want to ask comparative questions, relat­ work primarily on empirical questions, but ing one social context within a society to unless they are guided in research by some another or contrasting examples drawn knowledge of theoty, their work is unlikely from different societies. There are signifi- to be illuminating (see table 2.1).

Table 2.1 The sociologist's line of questioning Factual question What happened? Smce the 1980s, girls have been attaining better educational results in school than boys. Comparative question Dld this happen everywhere? V/as this a global phenomenon, or did it occur just in Britain, or only in a certain region of Britain? Developmental question Has this happened over time? 'What have been the patterns of girls' educational attainment over time? Theoretical question What underlies this phenomenon? Why are girls now performing better in school? What factors would we look at to explain this change? Asking and Answering Sociological Questions

At the same time, sOciologists strive not to attain theoretical knowledge for its own sake. A standard view is that while the soci­ ologist's values should not be permitted to bias their conclusions, social research should be relevant to real-world concerns. In this chapter, we look further into such issues by asking whether it is possible to produce 'objective' knowledge. We begin by stressing the scientific nature of sociology, before examining the stages involved in sociological research. Some of the most widely used research methods are then compared, as we consider some actual investigations. As we shall see, there are often significant differences between the way research should ideally be carried out and real-world studies. Sociological research, like other scientific research, is the art of the possible.

''''m cl s(l(";al sdl'ltlisl, Mich",,!. Thalll/e,ms I call" explail1 Is sociology scientific? I'/atri(ity or .lIIythiJlg like I h.l/, Imf if )'011 ('l'('r /(',1111 to kllOI(' ahlllll Pl'u!,//' 1'1/1 tlJ(' 11/(111. " As discussed in chapter 1, Auguste Co mte Cl The New Yorker Collection 1986 JB. Handelsman from cartoonbank.com. AD righls reserved. saw sociology as an emerging science, which should adopt the successful (posi­ tivist) methods of the natural sciences such ical world, so sociology and the natural as physics and chemistry. Durkheim, Marx sciences cannot be identical. Unlike objects and the other founders of sociology also in nature, humans are self-aware beings who thought of sociology as a scientific subject; confer meaning and purpose on what they but can we really study human social life in do. We cannot even describe social life accu­ a scientific way? Are Laud Humphreys' rately unless we first grasp the concepts that observations on the tearooms really people apply in their own behaviour. For scientific? Before we can answer, we must instance, to describe a death as a 'suicide' first understand what this word means; means knowing what the person in question what exactly is science? was intending when he died. Suicide can Science is the use of systematic methods only occur when an individual actively has of empirical investigation, the analysis at self-destruction in mind. If he accidentally data, theoretical thinking and the logical steps in front of a car and is killed, he cannot assessment of arguments to develop a body be said to have committed suicide. at knowledge about a particular subject­ The fact that we cannot study human matter. Sociology is a scientific endeavour, beings in exactly the same way as objects in according to this definition, because it nature is in some ways an advantage to soci­ involves systematic methods of empirical ology. Sociological researchers profit from investigation, the analysis of data and the being able to ask questions directly of those assessment of theories in the light of they study - other human beings - and get evidence and logical argument. responses that they understand. This oppor­ Studying human beings, however, is tunity to converse with the participants of different from observing events in the phys- research studies and confirm the researcher's ASKING AND ANSWERING SOCIOLOGICAL QU ESTIONS interpretations means that sociological findings are, at least potentially, even more Define the problem Select a topiC for research. reliable (different researchers would arrive at the same re sults) and valid (the research actually measures what it is supposed to) than those from the natural sciences. Review the literature Howeve r, in other respects, sociology creates Familiarize yourself with existing research on the topic . difficulties that are not encountered by natu­ ral scientists. People who are aware that thei r activities are being scrutinized may not behave in the same way that they normally Formulate a hypothesis do. They may consciously or unconsciously What do you intend to test? What is the relationship between the variables? portray themselves in a way that differs from their usual attitudes. They may even try to 'assist' the researcher by giving the responses they believe he or she wants. Researchers Select a research design studying the behaviour of, say, chemicals or Choose one or more research methods: experiment, survey, observation , use of frogs do not have this problem. existing sources.

THINKING CRITICALLY Reflecting on the discussion so far, in Carry out the research Collect your data , record information. what ways does sociology differ from the natural sciences, such as physics and chemistry? Were Laud Humphreys'

research methods 'scientific', for Interpret your results example? If we were allowed to use the Work oul the implications of the dala you collect. same methods to carry out a similar study today, would we be likely to get the same or different results? Thinking more generally and sociologically now, Report the research findings What is their significance? How do th ey how might recent social changes affect relate 10 previous findings? such a study and its potential findings?

Your findings are registered and The research process discussed in the wider academic community - leading perhaps to the Let us first look at the stages normally initiation of further research. involved in research work. The research process takes in a number of distinct steps, leading from when the investigation is Figure 2,1 Steps in the research process begun to the time its findings are published or made available in written form (see figure 2.1). knowledge about certain institutions, social processes or cultures. A researcher might set Defining the research problem out to answer questions like 'What propor­ All research starts from a research problem. tion of the population holds strong religious This is sometimes an area of factual igno­ beliefs?' 'Are people today really disaffected rance: we may simply wish to improve our with "big government"?' 'How far does the Asking and Answering Sociological Questions economic position of women lag behind What aspects of the problem has their that of men?'. research left unanalysed? Drawing upon The best sociological research, however, others' ideas helps the sociologist to clarify begins with problems that are also puzzles. the issues that might be raised and the A puzzle is not just a lack of information, but methods that might be used in the research. a gap in our understanding. Much of the skill in producing worthwhile sociological Making the problem preCise research consists in correctly identifying A third stage involves working out a clear puzzles. Rather than simply answering the formulation of the research problem. lfrele­ question 'What is going on here?', puzzle­ vant literature already exists, the researcher solving research tries to contribute to our might return from the library with a good understanding of why events happen as notion of how the problem should be they do. Thus we might ask: why are approached. Hunches about the narure of patterns of religious belief changing? What the problem can sometimes be turned into accounts for the decline in the proportions a definite hypothesis. Although rooted in an of the population voting in elections in educated guess about what is going on, a recent years? Why are women poorly repre­ hypothesis clearly states this in exact sented in high-status jobs? language so that it can be tested. If the No piece of research stands alone. research is to be effective, hypotheses must Research problems come up as part of be formulated in such a way that the factual ongoing work; one research project may material gathered will provide evidence that easily lead to another because it raises either supports or disproves them. issues the researcher had not previously considered. A sociologist may discover Working out a design puzzles by reading the work of other The researcher must then decide just how researchers in books and professional jour­ the research materials are to be collected. A nals or by being aware of specific trends in range of different research methods exists, society. For example, over recent years, and which one is chosen depends on the there have been an increasing number of overall objectives of the study, as well as the programmes that seek to treat the mentally aspects of behaviour to be analysed. For ill within the community, rather than some purposes, a survey (in which ques­ confining them in asylums. SOCiologists tionnaires are normally used) might be suit­ might be prompted to ask: what has given able. In other circumstances, interviews or rise to this shift in attitude towards the an observational study, such as that carried mentally ill? What are the likely conse­ out by Laud Humphreys, might be more quences both for the patients themselves appropriate. We shall learn more about and for the rest of the community? various research methods later in this chapter. Reviewing the evidence Once the problem is identified, the next step Carrying out the research taken in the research process is usually to At the pOint of actually proceeding with the review the available evidence in the field; it research, unforeseen practical difficulties might be that previous research has already can easily crop up. It might prove impossi­ satisfactorily clarified the problem. If not, ble to contact some ofthose to whom ques­ the sociologist will need to sift through tionnaires are to be sent or those people the whatever related research does exist, to see researcher wishes to interview. A business how useful it is for his or her purpose. Have firm or government agency may be unwill­ previous researchers spotted the same ing to let the researcher carry out the work puzzle? How have they tried to resolve it? planned. Difficulties such as these could ASK ING AND ANSWERING SOC IOLOGICAL QUESTIONS

TableZ.Z Number of motor vehicles per 1,000 inhabitants: selected countries 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Austria 462 463 503 515 528 543 495 509 529 544 555 565 537 Belgium 432 442 441 454 464 487 494 482 490 500 511 517 520 Canada 600 619 627 595 569 565 565 564 580 566 569 572 581 Germany 527 527 427 478 523 540 547 551 556 564 570 582 589 Greece 248 246 257 271 283 298 313 328 351 378 406 428 450 Portugal 310 370 407 439 438 501 533 569 610 654 698 711 756 Thrkey 57 47 53 61 64 68 97 105 111 116 124 148 148 UK 443 433 453 441 439 428 448 458 474 486 493 516 533 United States 842 718 779 725 719 771 783 784 792 798 810 816 807

Source ' OECD Factbook 2005

2.1 Reading and interpreting tables

You will often come across tables when on findings originally reported elsewhere, a reading sociological literature. They source will be included. The source sometimes sometimes look complex, but are easy to gives you some insight into how reliable the decipher if you follow a few basic steps, listed information is likely to be, as well as showing below; with practice, these will become where to find the original data. In table 2.2, the automatic. Do not succwnb to the temptation to source note makes clear that the data have been skip over tables; they contain information in taken from a publication by the OECD concentrated form, which can be read more 3. Read the headings along the top and left-hand quickly than would be possible if the same side ofthe table. (Sometimes tables are material were expressed in words. By arranged with 'headings' at the foot rather than becoming skilled in the interpretation of tables, the top.) These tell you what type of information you will also be able to check how far the is contained in each row and column. In reading conclusions drawn by a writer actually seem the table, keep in mind each set of headings as justified. you scan the figures. In our example, the headings on the left give the countries involved, 1. Read the title in full. Tables frequently have while those at the top refer to the levels of car longish titles, which represent an attempt by ownership and the years for which they are the researcher to state accurately the nature given. of the information conveyed. The title of 4. Identify the units used; the figures in the body of table 2.2 gives first the subject of the data, the table may represent cases, percentages, second the fact that the table provides averages or other measures. Sometimes it may material for comparison, and third the fact be helpful to convert the figures to a form more that data are given only for a hrnited useful to you: if percentages are not provided, nwnber of countries. for example, it may be worth calculating them. 2. Look for explanatory comments, or notes, 5. Consider the conclusions that might be about the data. Notes may say how the reached from the information in the table. Most material was collected, or why it is tables are discussed by the author, and what he displayed in a particular way. Many of the or she has to say should of course be borne in tables used throughout this book contain mind. But you should also ask what further explanatory notes. If the data have not been issues or questions could be suggested by the gathered by the researcher but are based data. Asking and Answering Sociological Questions

Several interesting trends can be seen in the ovmership ratios as a rough indicator of differential figures in our table. First, the level of car ownership prosperity. Third, in nearly all countries varies consIderably between different countries. represented, the level of car ownership increased The number of cars per 1,000 people was more between 1990 and 2002, but in some the rate of than five times greater in the United States than in increase has been higher than in others - probably Thrkey in 2002. Second, there is a clear connection indicating differences in the degree to which between car ownership and the level of affluence of cotmtries have successfully generated economic a country. In fact, we could probably use car growth or are catching up.

DOONESBURY by Garry Trudeau

ACJ/J4J.J,.y, MIf(.S , THAr'S~- 1HIM; (Y A MIS~ . INA~P()tJ..9J%ty:. me I?IiSf'Ot'.IlEN INPf(A7ElJ '!HAT ff)U5 HA/) /'K) f;F­ f6CTCW W~ OPtNCWS. ~~';r..!!;~

potentially bias the result of the study and general rule the higher the proportion of give a false interpretation. For example, if non-responses in the sample. the more the researcher is studying how business likely it is that the survey of those who do corporations have complied with equal take part will be biased. Eve n if every opportunities programmes for women, attempt is made to reduce bias in the then companies that have not complied survey, the observations that sociologists may not want to be studied. Clearly, this make in carrying out a piece of research are could result in a systematic bias in the results. Bias can enter the research process in THINKING CRITICALLY many ways, For example, if a piece of Familiarize yourself again with table 2.2 research is based on surveys of participants' and reflect on what questions might views, it may be easy fo r the researcher to follow from this. For example, did you push the discussion in a particular way, find anything surprising here? How such as asking leading questions that follow could we find out why vehicle their own particular prejudices .

Second World Will, a strong correlation can be found between the decline in pipe­ smoking and the decrease in the number of people who regularly go to the cinema. Clearly one change does not cause the other, and we would find it difficult to discover even a remote causal connection between them. There are many instances, however, in which it is not so obvious that an observed correlation does not imply a causal relation­ ship. Such correlations are traps for the unwary and easily lead to questionable or false conclusions. In his classical work of 1897, Suicide (discussed in chapter 1), Emile Durkheim found a correlation between rates of suicide and the seasons ofthe year. In the societies that Durkheim studied, levels of suicide increased progressively from Janu­ ary to around June or July. From that time onward they declined over the remainder of the year. It might be supposed that this demonstrates that temperature or climatic change are causally related to the propensity of individuals to kill themselves. We might perhaps surmise that as temperatures increase, people become more impulsive and hot-headed, leading to higher suicide rates. However, the causal relationship here has nothing directly to do with temperature Sociologists might be interested in the or climate at all. In spring and summer, most reasons why some young children, but not people engage in a more intensive social life others, smoke. However, it can be cliJJicult to than they do in the winter months. lndividu­ establish a causal relationship between the als who are isolated or unhappy tend to different factors involved. experience an intensification of these fe el­ ings as the activity level of other people rises. instance, between level of educational Hence they are likely to experience acute achievement and occupational success in suicidal tendencies more in spring and modern societies. The better the grades an summer than they do in autumn and winter, individual gets in school, the better-paid the when the pace of social activity slackens. We job he is likely to ge t. What explains this always have to be on our guard both in correlation? Research tends to show that it is assessing wheth er correlation involves not mainly school experience itself; levels of causation and deciding in which direction school attainment are influenced much causal relations run. more by the type of home from which the person comes. Children from better-off Causal mechanisms homes, whose parents take a strong interest Working out the causal connections in their learning ski lls and where books are involved in co rrelations is often a difficult abundant, are more likely to do well than process. There is a strong correlation, for those coming from homes where these ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS

LUC'{, YOU'RE YOU CAN'T PROVE IN ALL PR06ABILlT't', WORST PLAYER IN T~AT! YOU SHOULD YOU ARE THE WOR5T HlE HISTORV OF NEVER SAnHINGS PLA,(ER IN HIE THE THAT YOU CAN'T 1115TOR,( OF THE GAME! G,';:"

qualities are lacking. The causal mecha­ To find out whether a correlation between nisms here are the attitudes of parents variables is a causal connection, we use towards their children, together with the controls, which means we hold some vari­ facilities for learning that a home provides. abies constant in order to look at the effects Causal connections in sociology should of others. By doing this, we are able to judge not be understood in too mechanical a way. between explanations of observed correla­ The attitudes people have and their subjec­ tions, separating causal from non-causal tive reasons for acting as they do are causal relationships. For example, researchers factors in relationships between variables in studying child development have claimed social life. that there is a causal connection between maternal deprivation in infancy and serious A discussion of some recenl 'critical personality problems in adulthood. realist' approaches to environmental (,Maternal deprivation' means that an sociology can be found in chapter 5, 'The Environment'. infant is separated from its mother fo r a long period - several months or more - during the early years of its life.) How might we test Controls whether there really is a causal relationship In assessing the cause or causes that explain between maternal deprivation and later a correlation, we need to distinguish inde­ personality disorders? We would do so by pendent variables from dependent vari­ trying to control, or 'screen out', other ables. An independent variable is one that possible influences that might explain the produces an effect on another variable. The correlation. variable affected is the dependent one. In One source of maternal deprivation is the the example just mentioned above, admission of a child to a hospital for a academic achievement is the independent lengthy period, during which it is separated variable and occupational income the from its parents. Is it attachment to the dependent variable. The distinction refers mother, however, that really matters? to the direction ofthe causal relation we are Perhaps if a child receives love and attention investigating. The same factor may be an from other people during infancy, she might independent variable in one study and a subsequently be a stable person. To investi­ dependent variable in another. It all gate these possible causal connections, we depends on what causal processes are being would have to compare cases where children analysed. Ifwe were looking at the effects of were deprived of regular care from anyone differences in occupational income on with other in stances in which children were lifestyles, occupational income would then separated from their mothers but received be the independent variable rather than the love and care from someone else. If the first dependent one. group developed severe personality difficul- Asking and Answering Sociological Questions ties but the second group did not, we would Other interpretations of the correlation are suspect that regular care from someone in possible. It has been proposed, for instance, infancy is what matters, regardless of that people who are predisposed to get lung whether or not it is the mother. In fact, chil­ cancer are also predisposed to smoke. In dren do seem to prosper normally as long as this view, it is not smoking that causes lung they have a loving, stable relationship with cancer, but rather some built-in biological someone looking after them; this person disposition to smoking and cancer. does not have to be the mother herself.

Identifying causes Research methods There are a large number of possible causes that could be invoked to explain any given A common distinction is often made in soci­ correlation. How can we ever be sure that we ology between quantitative and qualitative have covered them all? The answer is that we research methods and traditions; the cannot be sure. We would never be able to former is associated with functionalism and carry out and interpret the results of a piece positivism, the latter with interactionism of sociological research satisfactorily if we and the search for meanings and under­ were compelled to test for the possible standing. As the term suggests, quantitative influence of every causal factor we could methods try to measure social phenomena imagine as potentially relevant. Identifying and will use mathematical models and, causal relationships is normally guided by often, statistical analysis to explain them. previous research into the area in question. Qualitative methods, on the other hand, If we do not have some reasonable idea attempt to gather detailed, rich data allow­ beforehand of the causal mechanisms ing for an in-depth understanding of indi­ involved in a correlation, we would proba­ vidual action in the context of social life. As bly find it very difficult to discover what the a rough-and-ready guide to a diverse range real causal connections are. We would not of sociological research methods, this know what to test/or. distinction is a starting point and many A good example of how difficult it is to be sociologists will tend to specialize in, or sure of the causal relations involved in a even favour, one tradition rather than correlation is given by the long history of another. However, there is a danger that the studies of smoking and lung cancer. two traditions will be seen as opposing Research has consistently demonstrated a 'camps' with entirely different approaches strong correlation between the two. Smok­ to research. This would not be very ers are more likely to contract lung cancer productive. than non -smokers, and very heavy smokers In fact, many research projects today are more likely to do so than light smokers. make use of mixed methods - both quanti­ The correlation can also be expressed the tative and qualitative - in order to gain a other way around. A high proportion of more comprehensive understanding and those who have lung cancer are smokers or explanation of the subject being studied. have smoked for long periods in their past. The findings from separate quantitative and There have been so many studies confirm­ qualitative studies can also be combined. ing these correlations that it is generally For example, some feminist sociologists accepted that a causal link is involved, but favour qualitative methods, which, they the exact causal mechanisms are thus far argue, allow the authentic voices of women largely unknown. to be heard in ways that quantitative studies However much correlational work is done do not. But without quantitative studies it on any issue, there always remains some would not have been possible to measure doubt about possible causal relationships. the extent of gender inequalities in society ASK ING AND ANSWERI NG SOCIOLOGICAL QUESTIONS

Table 2.3 Four of the main methods used in sociological research Research method Strengths Limitations Fieldwork Usually generates richer and more Can only be used to study in-depth information than other methods. relatively small groups or or conununities. Ethnography can provide a broader Findings rrught only apply to the understanding of social processes. groups or cormnllIlities studied; it 15 not easy to generallze on the basis of a single fieldwork study. Surveys Make possible the efficient collection The material gathered may be of data on large numbers of individuals. superficial; where a questionnaire lS highly standardized, Important

differences between respondents I viewpoints may be glossed over. Allow for precise comparisons to be Responses may be what people made between the answers of profess to believe rather than respondents. what they actually believe.

Experiments The influence of specific variables Many aspects of SOCial life cannot can be controlled by the investigator. be brought into the laboratory. Are usually easier for subsequent The responses of those studied researchers to repeat. may be alIected by their experimental situation. Documentary Can provide source of in-depth materials The researcher is dependent on research as well as data on large numbers, the sources that exist, which may depending on the type of documents be partial. studied. Is often essential when a study is either The sources may be difficult to wholly histoncal or has a defined interpret in terms of how far they historical dimension. represent real tendencies, as in the case of some offiCial statistics.

or set those individual women's voices into interviews as his main research methods. their wider societal context. Sociologists Here, the investigator hangs out, works or have to be prepared to use the most lives with a group, organization or commu­ appropriate methods for the questions they nity and perhaps takes a direct part in their want to answer. activities. Where it is successful. ethnogra­ Next, we look at the various research phy provides information on the behaviour methods sociologists commonly employ in of people in groups, organizations and their work (see table 2.3). communities, and also on how those people understand their own behaviour. Ethnography Once we see how things look from inside a given group, we are likely to develop a Laud Humphreys used ethnography, a type better understanding not only of that group, of fieldwork, or first-hand study of people, but of social processes that transcend the using participant observation and/or situation under study. Ethnography is one Asking and Answering Sociological Questions 51

researcher in a focus group is more partici­ pant than detached observer and may well Tell Me .\?out the influence the responses of the group. There experience of livin~ in uni'\'i is therefore a danger that participants will with nqture, like your tnv.1 perform according to their perception of the .ncestors thous.nds of researcher's expectations. ye.rs \?e-\'ore you. In the traditional works of ethnography, accounts were presented without very much information about the researchers themselves. This was because it was believed that an ethnographer could pres­ ent objective pictures of the societies they studied. More recently. ethnographers have increasingly tended to talk about them· selves and the nature of their connection to the people under study. Sometimes, for example, it might be a matter of trying to consider how one's race, class or gender affected the work, or how the power differ­ ences between observer and observed distorted the dialogue between them. For a long while, it was usual for research based on participant observation to exclude any account ofthe hazards or problems that had to be overcome, but more recently the published reminiscences and diaries of of a number of qualitative research fieldworkers have been more open about methods used in sociology that aims to them. Frequently, feelings of loneliness gain an in·depth knowledge and under· must be coped with - it is not easy to fit into standing of relatively small·scale social a social context or community where you do phenomena. not really belong. The researcher may be In recent years, sociologists have made constantly frustrated because the members use of focus groups, which have previously of the group refuse to talk frankly about been widely used in marketing and surveys themselves; direct queries may be of political attitudes, as a qualitative welcomed in some contexts but met with a research method. Focus groups are essen­ chilly silence in others. Some types of field­ tially 'group interviews' in which a particu· work may even be physically dangerous; for lar group of people - usually between four instance, a researcher studying a delinquent and ten individuals - are gathered together gang might be seen as a police informer or to discuss a subject and exchange views. might become unwittingly embroiled in The researcher acts as a moderator but also conflicts with rival gangs. asks specific questions relating to the Ethnographic studies also have other research study, to direct the discussion. major limitations. Only fairly small groups Focus groups can increase the size of a or communities can be studied and much sample quite easily and because of their depends on the skill of the researcher in interactive nature, any possible misunder­ gaining the confidence of the individuals standings can be clarified, thereby increas· involved. Without this skill, the research is ing the validity of a study's findings. unlikely to get off the ground at all. The However, critics point out that the reverse is also possible. A researcher could ASK ING AND ANSWERING SOCIOLOGICAL QU ESTlONS

well, or even that two different researchers would come to the same conclusions when studying the same group. This is usually less of a problem in survey research. In a survey, questionnaires are either sent out or administered directly in interviews to a selected group of people - sometimes as many as several thousand. Sociologists refer to this group of people, whatever its size, as a population. Whilst ethnographic work is well suited to in -depth studies of small sli ces of social life, survey research tends to produce information that is less detailed but which can usually be applied over a broader area. Surveys are the most widely used type of quantitative research method, allowing social phenomena to be measured and then analysed using mathematical models and statistical techniques. Many government bodies and private polling agencies also make extensive use of surveys to gain knowledge of people's atti­ tudes and voting intentions. These may be conducted through face-to-face interviews, telephone calls, postal questionnaires and, increasingly, online, via the Internet and email. Whichever method is adopted, the great advantage of surveys is that they allow researchers to collect large amounts of comparable data, which can be manipu­ In fieldwork, sociologists have to become lated, usually using computer software, to close to the communities they are studying, find out whether there are any significant but not so close that they lose their outsider's correlations between variables. eye. Standardized, open-ended, semi­ structured questionnaires begin to identify so closely with the group Three types of questionnaire are used in that he or she becomes too much of an surveys. Some contain a standardized, or 'insider' and loses the perspective of an fixed-choice, set of questions, to which only outside observer. a fixed range of responses is possible - for instance, ' Yes / No / Don't know' or' Very likely Surveys / Likely / Unlikely / Very unlikely'. Such surveys have the advantage that responses Interpreting field studies usually involves are easy to count and compare, since only a problems of generalization. Since only a small number of categories are involved. On small number of people are under study, we the other hand, because they do not allow cannot be sure that what is found in one for subtleties of opinion or verbal expres­ context will apply in other situations as sion, the information they yield is likely to Asking and Answering Sociological Questions be restricted in scope and sometimes 'What is your marital status?' might baffle misleading. some people. It would be more appropriate Other questionnaires are open-ended: to ask, J\re you single, married, separated, or respondents have more opportunity to divorced?' Most surveys are preceded by express their views in their own words and pilot studies in order to pick up problems are not limited to making fixed -choice not anticipated by the investigator. A pilot responses. Open-ended questionnaires study is a trial run in which just a few people typically provide more detailed information complete a questionnaire. Any difficulties than standardized ones. The researcher can can then be ironed out before the main follow up answers to probe more deeply survey is done. into what the respondent thinks. On the other hand, the lack of standardization Sampling means that responses may be more difficult Often sociologists are interested in the char­ to compare statistically. acteristics oflarge numbers of individuals ­ A very popular and widely used compro­ for example, the political attitudes of the mise between these two is the semi­ British population as a whole. It would be structured interview questionnaire, which impossible to study all these people directly, presents some standardized questions - the so in such situations researchers engage in data from which may be analysed statisti­ sampling - they concentrate on a sample, or cally later - but also includes interview small proportion, of the overall group. One prompts for more in-depth answers and can usually be confident that results from a sometimes allows interviewees to stray from population sample, as long as it was prop­ the schedule when necessary. Semi-struc­ erly chosen, can be generalized to the total tured interview schedules tend to pursue population. Studies of only two to three relevant research themes rather than highly thousand voters, for instance, can give a specific, researcher-defined questions. very accurate indication ofthe attitudes and Questionnaire items are normally listed voting intentions of the entire population. so that a team of interviewers can ask the But to achieve such accuracy, a sample must questions and record responses in the same be representative: the group of individuals predetermined order. All the items must be studied must be typical of the population as readily understandable to interviewers and a whole. Representative sampling is more interviewees alike. In the large national complex than it might seem, and statisti­ surveys undertaken regularly by govern­ cians have developed rules for working out ment agencies and research organizations, the correct size and nature of samples. interviews are carried out more or less A particularly important procedure used simultaneously across the whole country. to ensure that a sample is representative is Those who conduct the interviews and random sampling, in which a sample is those who analyse the results could not do chosen so that every member of the popula­ their work effectively if they constantly had tion has the same probability of being to check with each other about ambiguities included. The most sophisticated way of in the questions or answers. obtaining a random sample is to give each Questionnaires should also take into member of the population a number and consideration the characteristics of respon­ then use a computer to generate a random dents. Will they see the point the researcher list, from which tlle sample is derived - for has in mind in asking a particular question? instance, by picking every tenth number. Have they enough information to answer One famous early example of survey usefully? Will they answer at all? The terms research was called, 'The People's Choice?', of a questionnaire might be unfamiliar to a study carried out by Paul Lazarsfeld the respondents. For instance, the question and colleagues more than 60 years ago 54 ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS

(Lazarsfeld et al. 1948). This study, which misusers or people who self-harm, for investigated the voting intentions of resi­ example), it may be the only practical way of dents of Erie County, Ohio, during the 1940 gathering an adequate sample. Simi larly, campaign for the US presidency, pioneered snowball sampling, in which existing parti­ several of the main techniques of survey cipants are used to recruit other partici­ research in use to this day. In order to probe pants (usually via their own network of a little more deeply than a single question­ contacts and friends) is a tried and tested naire could do, the team interviewed each method of gaining access to a larger sample member of a sample of voters on seven than would otherwise be the case. separate occasions. The aim was to trace and understand the reasons for changes in Advantages and disadvantages of voting attitudes. surveys The research had a number of hypotheses Surveys are widely used in sociological to test. One was that relationships and research, for several reasons. Responses to events close to voters in a community can questionnaires can be more easily quanti­ influence voting intentions more than fied and analysed than material generated distant world affairs, and the findings on the by most other research methods; large whole confirmed this. The researchers numbers of people can be studied; and, developed sophisticated measurement given sufficient funds, researchers can techniques for analysing political attitudes; employ an agency specializing in survey yet their work also made significant contri­ work to collect the responses. The butions to theoretical thinking. The study scientific method is the model for this kind showed that some individuals - opinion of research, as surveys give researchers a leaders - tend to shape the political opin­ statistical measure of what they are ions of those around them. People's views studying. are not formed in a direct fashion, but in a Many sociologists today, however, are two-step process. In the first step, opinion critical of the survey method. They argue leaders react to political events; in the that an appearance of precision can be given second step, those leaders influence people to findings whose accuracy may be dubious, around them - relatives, friends and given the relatively shallow nature of most colleagues. The views expressed by opinion survey responses. Levels of non-response leaders, filtered through personal relation­ are sometimes high, especially when ques­ ships, then influence the responses of other tionnaires are sent and returned through individuals towards political issues of the the mail. It is not uncommon for studies to day. This study shows that good survey be published, based on results derived from research does more than simply describe little over half of those in a sample, although social phenomena, it can also aid our theo­ normally an effort is made to re-contact retical understanding. non-respondents or to substitute other There are other types of sampling that are people. Little is known about those who used by sociologists. In some types of choose not to respond to surveys or refuse research, it may be necessary to use conven­ to be interviewed, but survey research is ience sampling. This means taking your often experienced as intrusive and time­ sample from wherever you can! Because consuming. convenience sampling is less systematic and rigorous than other types, the results it Experiments generates have to be treated with caution. Nonetheless, in applied research or studies An experiment can be defined as an of hard-to-reach social groups who may be attempt to test a hypothesis under highly reluctant to come forward (substance controlled conditions established by an Asking and Answering Sociological Questions

2.2 Statistical terminolo!JV

Research in sociology often makes use of sununarize this data that most of the people o\'m far statistical techniques in the analysis of findings. more than they actually do. Some are highly sophisticated and complex, In such instances, one of two other measures but those most often used are easy to may be used. The mode is the figure that occurs understand. The most common are measures most frequently in a given set of data. In our of central tendency (ways of calculating example, it is £40,000. The problem with the mode averages) and correlation coefficients is that it does not take into account the overall (measures of the degree to which one variable distribution of the data - i.e., the range of figures relates consistently to another). covered. The most frequently occurring case ID a There are three methods of calculating set of figures is not necessarily representative of averages, each of which has certain their distribution as a whole and thus may not be a advantages and shortcomings. Take as an useful average. In this case, £40,000 is too close to example the amount of personal wealth the lower end of the figures. (including all assets such as houses, cars, bank The third measure is the median, which is the accounts and investments) owned by 13 middle of any set of figures; here, this would be the individuals. Suppose they own the following seventh figure, again £40,000. Our example gives amounts: an odd number of figures: 13. If there had been an even number - for instance, 12 - the median would 1. £000 (zero) be calculated by taking the mean of the two middle 2 £5,000 cases, figures 6 and 7. Like the mode, the median 3. £10,000 gives no idea of the actual range of the data 4. £20,000 measmed. 5. £40,000 Sometimes a researcher Vvill use more than one 6. £40,000 measure of central tendency to avoid giving a 7. £40,000 deceptive picture of the average. More often, he 8. £80,000 will calculate the standard deviation for the data in 9. £100,000 question. This is a way of calculating the degree of 10. £150,000 dispersal, or the range, of a set of figures - which 11. £200,000 in this case goes from zero to £10,000,000. 12. £400,000 Correlation coefficients offer a useful way of 13. £10,000,000 expressing how closely cormected two (or more) variables are. VVhere two variables correlate The mean corresponds to the average, arrived completely, we can speak of a perfect positive at by adding together the personal wealth of all correlation, expressed as 1.0. VVhere no relation is 13 people and dividing the result by 13 . The found between two variables - they have no total is £11,085,000; dividing this by 13, we consistent cormection at all- the coefficient is zero. reach a mean of £852,692.31. This mean is often A perfect negative correlation, expressed as 21 .0, a useful calculation because it is based on the exists when two variables are in a completely whole range of data provided. However, it can inverse relation to one another. Perieet correlations be misleading where one or a small number of are never found in the social sciences. Correlations cases are very different from the majority In the of the order of 0.6 or more, whether positive or above exanlple, the mean is not in fact an negative, are usually regarded as indicating a appropriate measure of central tendency, strong degree of cormection between whatever because the presence of one very large figure, variables are being analysed. Positive correlations £ 10,000 ,000. skews the picture. One might get on this level might be found between, say, social the impression when using the mean to class background and voting behaviour. ASK I NG AND ANSWERING SOCIOLOGICAL QUESTIONS

Classic Studies 2.1 The social psychology of prison life

The research problem Most people have not experienced life in prison and find it hard to imagine how they would cope 'inside'. How would you fare? Could you work as a prison officer? What kind of a prison officer would you be? A disciplinarian maybe? Or perhaps you would adopt a more humanitarian approach to your prisoners? In 1971, a research team led by Philip Zimbardo decided to try and find out what impact prison would have on 'ordinary people'. In a study funded by the US Navy, Zimbardo set out to test the 'dispositional hypothesis', which dominated within the armed forces. This hypothesis suggested that constant conflicts between prisoners and guards were due to the individual characters of the guards and inmates - their personal dispositions. Zimbardo thought this may be wrong and set up an experimental prison to find out.

Zimbardo's explanation Zimbardo's research team set up a fake jail at Stanford Universi~ advertised for male volunteers to participate in a study of prison life and selected 24 mainly middle-class students Reaction to the mock prison regime during the who did not know each other prior to the Stanford prison experiment led to one inmate experiment. Each participant was then randomly staging a hunger strike to be let out. assigned as guard or prisoner. Following a standard induction process which involved being stripped naked, de-loused and because of the distress of the participants. Even photographed, also naked, prisoners stayed in before this, five 'prisoners' were released jail 24 hours a day; but guards worked shifts and because of extreme anxiety and emotional went home in between times. Standardized problems. Many 'guards', though, were unhappy uniforms were used for both roles. The aim was that the study had ended prematurely to see how playing these different roles would On the basis of the findings, Zimbardo lead to changes in attitude and behaviour. The concluded that the dispositional hypothesis results shocked the investigators. could not account for the participants' reactions. Students who played at being guards quickly Instead, he proposed an alternative 'situational' assumed an authoritarian manner; they explanation - behaviour in prisons is influenced displayed real hostility towards the prisoners, by the prison situation itself, not by the individual ordering them around and verbally abusing and characteristics of those involved. In particular, bullying them The prisoners, by contrast, the expectations attached to the roles being showed a mixture of apathy and rebelliousness - played tended to shape behaviour. Some of the a response often noted among inmates in real guards' behaviour had deteriorated - they prisons. These effects were so marked and the treated prisoners badly; regularly handing out level of tension so high that the 14-day punishments and appearing to take pleasure in experiment had to be called off after just 6 days the distress of the prisoners. Zimbardo suggests Aski ng and Answering Sociological Questions

that this is due to the power relationships the jail unrepresentative sample. A final criticism is that had established. Their control over prisoners' the constructed situation may invalidate the lives very quickly became a source of enjoyment findings. For example, participants knew their for the guards. On the other hand, following a imprisonment would only last 14 days and they short period of rebelliousness, prisoners very were paid $15 a day for their participation. rapidly exhibited a 'learned helplessness' and Established problems of prisons such as racism dependency As researchers, the study tells us and involuntary homosexuality were also absent. something important about why social Critics say that the experiment is therefore not a relationships very often deteriorate within meaningful comparison with real prison life. prisons and, by implication, other 'total institutions' (Goffman 1968 [1961]). This has little Contemporary significance to do with individual personalities and much In spite of the somewhat artificial situation - it more to do with the prison environment and the was an experiment, after all - Zimbardo's social roles within it. findings have been widely referred to since the 1970s. For example, Zygmunt Bauman's (1989) Critical points Modernity and the Holocaust draws on this study Critics argue that there were real ethical to help explain the behaviour of inmates and problems with the study Participants were not guards in Nazi-run concentration camps in the given full information about the purpose of the Second World War. Just as importantly; the research and it is therefore questionable general thesis emerging from the research - that whether they could really have given 'informed institutional settings can shape social relations consent'. Should the study even have been and behaviour - continues to inform allowed to go ahead? The sample selected was contemporary sociological studies, such as clearly not representative of the population as a those investigating care homes for older people, whole; all were students and male. Generalizing residential homes for disabled people and many about the effects of 'prison life' would therefore more. be very difficult with such a small and

investigator. Experiments are often used in small groups of individuals can be brought the natural sciences, as they offer major into a laboratory setting, and in such exper­ advantages over other research procedures. iments people know that they are being In an experimental situation the researcher studied and may not behave normally. Such directly controls the circumstances being studied. Psychologists studying individual behaviour also use laboratory-based exper­ THINKING CRITICALLY imentation extensively. However, compared What is your first response to the wi th these disciplines, the scope fo r experi­ Zimbardo study - can such an mentati on in sociology is quite restricted. e xperimental situation really reproduce Most sociological studies, even those of the authentic experie nce of a prison? individual actions, look to investigate the Which aspects of the prison experience relationship between m icro~ and macro so­ could an experiment never replicate? cial phenomena. To remove individuals Thinking more critically now, should from their social context for the purposes of social scientists be allowed to 'experiment' on human be ings at all? If experimentation effectively prevents this. not, does that mean there are things we Sometimes, sociologists may want to will just never know about? If they explore group dynamics - the way individu­ should, what limits should there be on als behave when in groups - and experi­ such experime nts? ments may then be possible. However, only 58 ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS

changes in subject behaviour are referred to the interviews, letters and newspaper arti­ as the 'Hawthorne effect'. In the 1930s, cles they collected. researchers conducting a work-productivity Other procedures of research do not study at the Western Electric Company's usually yield as much information as the Hawthorne plant near Chicago found to life-history method does about the devel­ their surprise that worker productivity opment of beliefs and attitudes over time. continued to rise regardless of which exper­ Life-historical studies rarely rely wholly on imental conditions were imposed (levels of people's memories, however. No rmally, lighting, break patterns, work team size and sources such as letters, contemporary so forth). The workers were conscious of reports and newspaper descriptions are being unde r scrutiny and accelerated used to expand on and check the valid­ their natural work pace. Nevertheless, as ity of the information that individuals 'Classic Studies 2.1' shows, we can still learn provide. things about social life from small-scale Sociologists' views differ on the value of experiments. biographical methods. Some feel they are too unreliable to provide useful informa­ Biographical research tion, but others believe they offer sources of insight that few other research methods can In contrast to experiments, biographical match. Indeed, some sociologists have research belongs purely to sociology and begun to offer refl ections on their own lives the other social sciences; it has no place in within their research studies, as a way of natural science. Biographical research has offering insights into the origins and devel­ become much more popular and widely opment of their theoretical assumptions used in sociology in recent decades and (see Mouzelis 1995). includes oral histories, narratives, autobi­ ographies, biographies and life histories Comparative research (Bryman 2008). These methods are used to explore how indivi duals experience social The research methods described so far are life and periods of social change, and how generally applied in a comparative context. they interpret their relationships with Comparative research is of central impor­ others in the context of a changing world. In tance in sociology, because making this way, biographical methods allow new comparisons allows us to clarify what is voices to enter sociological research; life going on in a particular area of research. Let histories are an example of such research us take the rate of divorce in many devel­ methods. oped societies- - i. e. the number of divorces Life hi stories consist of biographical granted each year - as an example. In the material assembled about particular indi­ early 1960s there were fewer than 30,000 viduals - usually as recalled by the individu­ divorces per year in the UK; by the early als themselve s. Life histories have been 1980s this figure had risen to around successfully employed in studies of major 160,000 or more. Do these changes reflect importance. One celebrated early study was specifi c features of British society? We can The Polish Peasant in Europe and America, find out by comparing divorce rates in the by W. I. Thomas and Florian Znaniecki, the UK with those of other countries. Such a five volumes of which were first published comparison reveal s that compared to most between 1918 and 1920 (Thomas and other Western societies the overall trends Znaniecki 1966). Thomas and Znaniecki are similar. A majority of Western co untries were ab le to provide a more sensitive and have experienced steadily climbing divorce subtle account of the experi ence of migra­ rates over the past half century. tion than would have been possible without Asking and Answering Sociological Questions 59

Historical analysis the First World Wa r (As hworth 1980). Ashworth was concerned with analys ing As discussed in chapter 1, a historical what life was like for men who had to endure perspective is often essential in sociological being under constant fire, crammed in close research. For we frequently need a time proximity for weeks on end. He drew on a perspective to make sense of the material we diversity of documentary sources: official coll ect about a particular problem. histories of the war, including those written Sociologists commonly want to investi­ about different military di visions and gate past events directly. Some periods of battalions; official publications of the time; history can be studied in a direct way while the notes and records kept informally by there are still survivors around - such as in individual soldiers; and personal accounts the case of the Holocaust in Europe during of war experiences. By drawing on such a the Second World War. Re search in oral variety of sources, Ashworth was able to history means interviewing people about develop a rich and detailed description of events they witnessed at some point earlier li fe in the trenches. He discovered that most in their lives. This kind of research work, soldiers formed their own ideas about how obviously, can only stretch at the most some often they intended to engage in combat 60 or 70 years back in time. with the enemy and often effectively Fo r historical re search on an earlier ignored the commands of their officers. For period, sociologists use documentary example, on Christmas Day, German and research from written record s, often Allied soldiers suspended hostilities, and in contained in the special coll ections of one place the two sides even staged an libraries or archives. The range of useful informal soccer match. documents is extensive, taking in personal sources such as diaries; official sources Comparative-historical research such as policy documents, records of births and deaths and tax records; documents Ashworth's research concentrated on a rela­ from private bodies like businesses and tively short time period, but there have been voluntary organizations; as well as maga­ many studies that have investigated social zines and newspapers. Depending on the change over much longer periods and research question. historical documents appli ed comparative research in that histor­ such as these can all constitute primary ical context. One of the more recent modern sources, just as much as the data recorded classics of comparative histori cal sociol­ in interviews with war survivors. However, ogy is Theda Skocpol's (1979) analys is of historical sociologists also make much use social revolutions, discussed in 'Classic of secondary sources: accounts of histori­ Studies 2.2 '. cal events written by people afterwards. Most documentary studies make use of both types. However, sociologists face the Sociological research in same issues as historians when they make the real world use of such sources. How authentic are the documents? Is the information within them All research methods, as was stressed reliable? Do they represent only a partial earlier, have their advantages and limita­ viewpoint? Documentary re search requires tions. Hence, it is common to combine a patient, systematic approach to sources several methods in a single piece of and their interpretation. research, using each to supplement and An interesting example of the use of check on the others, in a process known as histori cal documents is sociologist Anthony triangulation. We can see the value of Ashworth's study of trench warfare during combining methods - and, more generally, ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS

Classic Studies Z.Z Theda Skocpol's comparison of social revolutions I

I The research problem Skocpol's explanation ,I As all students of sociology and history are Skocpollooked at the processes of revolution in taught. the French Revolution of 1789 three different historical contexts: the 1789 transformed France forever But why did it French Revolution (1786-1800): the 1917 happen then and not at some other time? Was it revolutions in Russia (1917-21) and the an historical accident or was it inevitable? The revolutionary period in China (1911-49). Given early twentieth-century revolutions in China and the essentially historical questions she asked, Russia not only turned those countries into her main method was the use and careful communist societies, but significantly shaped the interpretation of a range of primary and direction of the modern world itself. Again, why secondary documentary sources. Although did they happen then? What caused them? The there are many differences between the three American sociologist Theda Skocpol (1947- ), cases of revolution, Skocpol argues that their set out to address these questions and to underlying structural causes are in fact similar. uncover the similarities and differences between She rejects the Marxist idea that revolutions are them. Skocpol set herself an ambitious task: to the product of mass (class-based) movements produce a theory of the origins and nature of with deep grievances. Instead, she agrees with revolution grounded in detailed empirical study the thesis that, 'revolutions are not made, they The result was a book, published as States and come'. That is, social revolutions are largely the Soclai RevoiutJODS (1979), that is now one of the result of the unintended consequences of classic studies of long-term social intentional human action. Before the Russian transformation. Revolution, for instance, various political groups

Social unrest does not necessarily lead to revolution. It is unclear what the outcome of protests in Tibet against Chinese rule will be. Asking and Answering Sociological Questions

were trying to overthrow the pre-existing Could things have turned out differently if regime, but none of these - including the individual actors and groups had chosen Bolsheviks, who eventually came to power - alternative courses of action? Are individuals so anticipated the revolution that occurred. A series powerless to influence change in the face of of clashes and confrontations gave rise to a structural pressures? A further criticism is of process of social transformation much deeper Skocpol's notion of 'cause' in this context. Some and more radical than anyone had foreseen. have argued that what her argument amounts to Skoepol's explanation is that all three is really a set of sophisticated generalizations in revolutions occurred in predominantly agrarian relation to the cases she studied. And although societies and were made possible only when the such generalizations work quite well for these existing state structures (administrative and specific cases. this is not the same thing as a military) were breaking down as they came general causal theory of social revolutions. So, under intense competitive pressure from other critics say, despite setting out to discover the states. In this context, it was peasant revolts and underlying causes and nature of social mass mobilizations that brought about social revolutions, in the end, Skocpol's study showed revolutions in France, China and Russia. Thus that each revolution has to be studied in its own Skocpol argued against the widespread notion right. that peasants were not a 'revolutionary class'. Some similarities with other revolutions in Contemporary significance Vietnam, Cuba, Mexico and Yugoslavia can also Skocpol's study has become a modern classic be seen. Skocpol's causal explanation focuses for two reasons. First, it developed a powerful on state structures; as these began to break causal explanation of revolutionary change, down, a power vacuum was created and states which emphasized the underlying social lost their legitimacy, enabling revolutionary structural conditions of revolution. Such a strong forces to take power. central thesis was, nevertheless, underpinned by Skocpol's research makes use of the 'logic of very detailed analysis of primary and secondary scientific experiment' for comparative studies, documentary sources. Hence, Skocpol outlined by John Stuart Mill in the mid-nineteenth successfully demonstrated that comparative­ century Skocpol adopts Mill's 'method of historical sociology could combine the study of similarity', taking three similar events large-scale, long-tenn social change with the (revolutions) in very different national contexts. empirical investigation of historical events' on This allows her to look for possible key the ground'. In essence, she brought together similarities across the three cases which can be the macro- and rnicrosociological aspects into identified as independent variables and thus, one theoretical framework Second, Skocpol help to explain the causes of political revolutions. made a very significant contribution to our understanding of revolutions. She showed that Critical points there are enough similarities across different Some of Skocpol's critics have raised questions revolutions to warrant pursuing general theories about her thesis's structural argument. This, they of social change. In this way, her thesis helped to say, leaves little room for active agency on the bridge the gap between mainstream historical part of people. How did peasant groups revolt? studies and the sociology of revolutions. Did leaders not play a part in the revolutions?

the problems and pitfalls of real sociological for him to find this out, because all he could research - by looking once again at Laud really do in the toilets was observe. The Humphreys' Tearoom Trade. norm of silence in the toilets made it One of the questions that Humphreys difficult to ask any questions, or even to talk. wanted to answer was: 'What kind of men In addition, it would have been very odd if came to the tearooms?' But it was very hard he had begun to ask personal questions of 62 ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS

people who basically wanted to be anony­ already know? Is itsimplythe tedious defini­ mous. tion of social phenomena with which we are Humphreys' solution was to try to find already familiar' Sociology at its worst can out more about the men in the 'tearooms' be all of these things, but it is never appro­ using survey methods. Standing by the door priate to judge any discipline by what its of the toilets, he would write down the car worst practitioners do. In fact, good sociol­ number-plates of people who pulled up to ogy sharpens our understanding of the the car park and went into the toilets for sex. obvious (Berger 1963) or it completely He then gave the numbers to a friend transforms our com.mon sense. In either who worked at the Department of Motor event, good sociology is neither tedious nor Vehicles, securing the addresses of the men. a restatement o[the obvious. Discussions of Months later, Washington University in St a new topic in this text sometimes begin Louis, in the United States, where Laud was with definitions of things that you may working, was conducting a door-to-door already understand. It is necessary for any survey of sexual habits. Humphreys asked academic discipline to define its terms. the principal investigators in that survey if However, when we define a family, for he could add the names and addresses of example, as a unit of people who are related his sample of tearoom participants. to one another, we do so not as an endpoint Humphreys then disguised himself as one but instead as a beginning. Without defining of the investigators and went to interview OllI terms, we often cannot move forward to these men at their homes, supposedly just sharper levels of understanding later on - to ask only the survey questions but actually good sociology never defines terms as an also to learn more about their social back­ end in itself. grounds and lives. He found that most of these men were married and otherwise led The influence of sociology very conventional lives. He often inter­ viewed wives and other family members as Sociological research is rarely of interest well. only to the intellectual community of soci­ ologists. For a start, much sociological THINKING CRITICALLY research funding comes from government sources and is directly linked to social issues Do you think Zimbardo's project would and problems. Many studies of crime and be allowed today? On reflection, should deviance, for example, target specific such projects be allowed in sociology at all? Considering Laud Humphreys' offences or types of offender with a view to gaining a better knowledge so that the social study of tearooms and Zimbardo's prison experiment, can the knowledge problems associated with crime can be benefits of a research study ever justify tackled more effectively. Sociologists also compromising a researcher's own work with voluntary agencies, public bodies ethical position? and businesses, bringing their research skills to bear on matters of interest not just to the sociologist. A fair amount of this work Restating the obvious? is applied social research, which does not simply aim to produce better knowledge, Because sociologists often study things that but also seeks to inform interventions we have some personal experience of, one aimed at improving social life. A study of the can sometimes wonder if sociology is effects on children of parental alcohol use, merely 'a painful elaboration of the obvious' for example, may be interested in whether a (Wright 2000). Is sociology merely a restate­ particular treatment programme has any ment, in abstract jargon, of things we effect on reducing alcoholism. Sociology's Asking and Answering Sociological Questions 63

im pact is not restricted to policy-oriented logical investigation. A way of describing research, however. this phenomenon, using the technical The resul ts of sociological research are concepts of sociology, is to say that sociol­ often disseminated th roughout society. ogy stands in a 'refl exive relation' to the Sociology, it must be emphasized, is not just human beings whose behaviour is studied. the study of societies; it is a significant Reflexivity, as we shall see in chapter 3, element in the continuing life of those soci­ describes the interchange between socio ­ eti es. Take the transformations taking place logical research and human behaviour. We in relation to marriage, sexuality and the should not be surprised that sociological family (discussed in chapters 9 and 14). Few find in gs ofte n correlate closely with people do not have some knowledge of common sense. The reason is not simply these changes, as a result of the filtering that sociology comes up with fin dings down of sociological research. Our thinking we knew already; it is, rather, that sociologi­ and behaviour are affected by sociological cal research conti nually infl uences what knowledge in complex and often subtle our common -sense knowledge of society ways, thus reshaping the very fi eld of socio- actually is.

Summary points 4. In fieldwork or participant observation, the researcher spends lengthy periods of time I. Sociologists investigate social life by posing with a group or community being studied. distinct questions and trying to find the Survey research involve s sending or answers to these by syste matic rese arch. administering questionnaire s to samples of a These questions may be factual, comparative, larger population. Documentary research uses developmental or theoretical. printed materials, from archives or other 2. All rese arch begins from a research problem, resources, as a source for information. Other which interests or puzzle s the investigator. research methods include experime nts, Gaps in the existing literature, theoretical biographical methods, historical analysis and debates, or practical issues in the social world comparative research. may suggest research problems. There are a 5. Each of these various methods of research has numbe r of clear steps in the development of its limitations. For this reason, researchers will research strategies - although these are rarely often combine two or more methods in the ir followed exactly in actual research. work, each being used to check or 3. A causal relationship between two events or suppleme nt the material obtained from the situations is one in which one event or others. This process is called triangulation. situation brings about the other. This is more 6. Sociological re search ofte n poses ethical problematic than it seems at first. Causation dilemmas. These may arise either where must be distinguishe d from correlation, which deception is practised against those who are refers to the existence of a regular the subjects of the research, or where the relationship between two variables. A variable publication of research findings might can be differences in age, income, crime adversely affect the feelings or lives of those rates, etc. We need to also distinguish studied. There is no entirely satisfactory way independent variables from dependent to deal with these issue s, but all researchers variables. An independent variable is a have to be se nsitive to the dilemmas they variable that produces an effect on another. pose. Sociologists often use controls to ascertain a causal re lationship. ASKING AND ANSWERING SOCIOLOGICAL QUESTIONS

••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• • Bryman's Social Research Methods, 3rd edn • Further reading • (Oxford: Oxford University Press, 2008), • • • Books on research methods are very which has become very widely adopted by • numerous, especially those pitched at lecturers for their research methods courses. • • introductory level students. This selection is Tim May's Social Research: Issues, Methods and • • therefore merely the tip of a large iceberg of Process, 3rd edn (Buckingham: Open • • • • literature. University Press, 2003) is a similarly much­ • • • • used and reliable guide. • • Novice researchers need a text that is both • • • • informative and practical, so something like One other book worth looking at is Darrell • • • • to • • Judith Bell's Doing Your Research Project: A Huffs How Lie with Statistics (London: • • • • Guide for First-time Researchers in Education, Penguin Books Ltd, 1991), which is • • • • Health and Social Science, 4th edn apparently the best-selling statistics book • • • • (Buckingham: Open University Press, 2007) ever written (see J. M. Steele, 'Darrell Huff • • is a very good place to begin thinking about and fifty years of how to lie with statistics', • • and planning a research project. Similarly, Statistical Science, 20/3 (2005): 205- 9). This is • • Keith F. Punch's Introduction to Social probably because its irreverent tone makes it • • Research: Quantitative and Qualitative attractive to students. It is, however, an • • Approaches, 2nd edn (London: Sage excellent guide to the misuse of statistical • • Publications, 2005) does exactly what it says information in society, which carries a • • • and covers a lot of issues and debates. Gary serious message. • • D. Bouma and Rod Ling's The Research • Finally, a good dictionary is usually a good • Process, Fifth Edition (Melbourne: Oxford • investment, so Victor Jupp's The SAGE • University Press, 2005) is also an excellent • Dictionary ofSocial Research Methods • introduction to research methods. • (London: Sage Publications, 2006) is well • For something more sophisticated and worth consulting. • • comprehensive, you could then try Aian • •••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• Asking and Answering Sociological Questions

••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• • The UK Data Archive - a centre of expertise in • Internet links • relation to the collection and storage of data on • • Ipsos MORI - a merged company (lpsos UK and a variety of subjects: • • MORI) focusing on market research and social www.data-archive.ac.uk/ • • research: • CESSDA - Council of European Social Science • www.ipsos-mori.com/ • Data Archives - houses many social science data • Intute - social science web resources for educa­ archives from across Europe covering many tion and research: types of research: www.intute.ac. uk/ socialsciences / www.nsd.uib.no/cessda/home.html UK National Statistics Online, which includes lots of survey research, but other types as well: • • www.statistics.gov.uk/default.asp • ••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••