s1
hehe1234
G W ' 00d7 475 ';a,-* ~ ~ C . U ~ U ~ L \ , ' Z , ~ ~ ~ ~ ~ ~ ~ - @whAqs, 5- 2 J 4 TQ. k ~ . ? f r : r . ( ( 5 7 , ) 7
4 HOW Sociologists DO Research JAMES M. HENSLIN
Guesswork does not go very far in helping us to understand the social world. Our guesses, hunches, and the rdea that pas for curninon sense may o r may not be correct SDciologists m ~ u t gather data in such a way that what they report i s objecbve-presentir~g information that repre- sents what is mally "out tl~ere." To rb so, they rnust use methods that other researchers can reprat (refllicntel to check their findings. They also must tie their findings ~nto b t h theory ad what other researchers have already reported. In this overview ol' rnearch methods. Henslin outlines the procedures that sociologists use to gather data.
R e n k had never felt fear before-at least not iike this I t had begun as a vague feeling that something was out of place. Then shc felt it c w p up her spine, slowly tightening as it clawed its
upwards. NOW it w s like it fist pounding inside her skull. Ren& never went anywhere with strangen Hadn't her par-
ents hainmered that into her head since she was n child? And now, at 19, she w n ' t about to start breaking that nile
And yet here she was, in a rar with a stranger. H e seemed nice enough. And it wasn't as thoudl he were some strange guy nn the side of the road or an).thlng. She had rtiet George at Patricia's p q , and. . . .
R e n h had first been attracted by hls dark e es They seemed to lght up his entire Face when he smiled h d when he asked her to dance, Renee felt flattered He was u little older, a little more sure of himself than most of the guys she knew. Renbr. liked that: It was a sign of rnaturig.
As the evening wore o n and he continued to be attentive to her, it seemed natural to accept his offer to take her home.
But then they passed the turn to her dorm. She didn't un- derstand hizi mumbled reply ahout "getting something," And as he turned OR on the muntry m d , that clawing at the back oC her - neck had Ire n
h he at her, llir c p r allnost picrcsd tho clarhen. "It's time to pay, Babe," he said, as he clawed at her blouse.
Ren& won't talk about that rught. She doesn't want to recall anything that happened then
IN THIS PAPER we examine how swiologists do research. As we look at liow they gather data, we focus on this basic question: How can we gitller reliable information on rape--which is to say on both rapists and their victims?
Sociology and Common Sense
Common sense will give ur sonle information. From common sense (a kind of knowledge not based on formal wvestigation, but on ideas that we pick up from our groups, mixed with abstractio~ls from our own experiences) we h o w that her rape was a significant event in RellCe's Me. And from common sense we know that rape has ongoing effects. that i t can trigger fears and anxieties, and that i t can make women distrust men
It so happens that these ideas are true. But many other common-sense ideas, even though g h g l y obvious to us. are rwt true, md so tlTe need re- search to test the validity and aLturaq of our i d e s . For example, common sense also tells us that one reason men rape is the ~ v r J n g clohng that some women wear. And common sense tells US that men who rape are sexually de- prived. These ~arnmon-sense ideas, howevrr, are not on target. Researchers have found that men who r a p don't care what a woman i s wearing; most don't even care who the woman is. She is simply an object for their lust, drives for power and expIoitation, and, sometimts, frustration and anger. Researchers have also found that rapists may or may not be semally deprived-the same as with men who do not rape For eumple, many rapists have wives or girl friends with whom they have an ongoing sexual rehtionshp.
If it is neither provmrtiw c l o h n g nor sexual deprivation, then. what d o e s cause rape? And what effects does r a p have 011 victims? Phrasing the matter ths way-~rlstead of zsuming that we h a w d ~ e answers-not only opens up our mill& but dso undersmres the pressing 11e.d for soc~o~ogical research, the need to search for empirical findings that will t i e us completely out of the realm of guesswork and well beyond common sense.
Let us s e e now how sociologsts do their research. We shall look first at a research model, and then at the research methods used in sociolog.
A Research Model
As shown in Figure 4.1, eight basic steps are involved in social research. As you look at each of these steps, be aware that this is an ideal model. In some re- search these steps are collapsed, in others their order may be rearranged, while in still others one or more steps may be omitted.
How Sociologis~ Do Research / 33
v 2 DeRn~np the problem
I
; I 6. Collaci~ng the dala I
I- ' 1. 2. 3. 4. 5. 6.
Surveys Secondary k r n ~ t s Experiments Unobtrusive Perticipsnt I I anabsis Measures ObservaUon 1 4 I
7. Analyzing the results 4 I
6. Sharlng Me results I I Stimulates L - - - - - Ideas lor - - - - - - - + - - I
More resear& Figure 4.1 r he Research Model (Modi@tion ofFig. 2.2 iri 5rliorfcr 1989.)
1. SE1,ECTINC A TOPIC
The first step is to select a topic. What is it that you want to h o w more about? Many sociologists simply foilow their curiosity, their drive to know. They hr- come interested in a particular topic, and they pursue it. Sometimes swiolo- gists choose a topic simply because funds are available. At other times, some social problem, such as rape, has become a pressing issue and the sociolqpt wants to gather data that will help people better understand-and perhaps help solve it .
2. DEFINING TIIE PROBLEM The second step is to define the problem, to determine what p u w;ult to lean1 a b u t the topic. To develop a researchable question, you need to focus or) a specific area or problem. For example, you may want to detertnine the educa- tion and work experiences of rapists, or the average age of their victims.
3. REVlEWNG T H E LITERATURE The third step is to review the literature. Nol-rody wa11ts to rediscover the wheel. If the question has already been answered, you want to h o w that. In addition, a review of what has been written on the topic cm stir your ideas, help sharpen your questior~s, and help you wmmplish the next step.
4. FOHMULATINC A HYPOTHESIS
The fourth step is to formulate a h y p o t h s i s , a statement of what you expect to find according to prehctions from a theory. A hypothesis predicts a rela- tionship between or among uoriables (factors thought to be significant). For example, the statement "Men who are more socially isolated are more Likely to rape than are men who are more socially intepated" is example of a hy- pothesis. Hypotheses (the plural) need ~peratior~al definitions-that is, pre- cise ways to measure their concepts. In dus exannple, you would need opera- tional definitions for three concepts: social integration, social isolation, and rape.
5 . CHOOSING A RESEARCH METHOD
The nays by which sociologists collect data are called research methods. You need to seieb a methcd that will answer the questions you have formulated. (In the next section, beginning on page 36, 1 go into detail concerning the six reseilrdl m e t h d s used by sociologists.)
The next step is to gather the data. Great care needs to be taken to assure that the data are both v d d and reliable. Validity means the extent to whch the op- erational definitions measure what you intend to measure. In other words, do your definitions or measures of social isolation and integration really measure these concepts and not something else?
The concept of rape is not as simple to define (or operationalize) as it may seem. For example, there are various degrres of sexual assault. Look at Table 4.1, which depicts a variety of iotced sexual actitities. Deciding which of these constitute rape for the purposes of yvur researdl project is an example of the difficulties of developing operational definitions. Certainly not all o f these acts are rape-and, therefore. not all of those who did them are rapists.
&liability refers to the extent to which your measures and studies give consistent results. Inadequate operational definitions and sampling (covered later) will prevent reliabiliiy. For exan~ple, if your measure of rape is inade- quate, other researcherr will exclude wts that you included, and include acts that you excluded. In that case, how cat1 you compare the results?
Bow Sociologists Do Research I 35
TABLE 4. L Date Rape and Other Unwanted Sexual Actlvities Experienced by College Undergraduates
WOMEN WHO REPORTED T H I S MEN WHO HAD H A P P E N E D REPORTED THEY
TO T H E M HAD DONE T H I S Uh'WANTED SEXUAL ACTrnlW
Be kissed without tongue mntaEt 3.7 2.2 He hssed with tongue contact 12.3 0.7 H e touchedkissed her b r e z t s
through her clothes 21.7 7.3 He touchedhsed her breasts
under her clothes 22.6 13.1 He touched her genitals through
her clothes 28.8 15.3 He touched her genitals under her
clothes 28.1 13.9 He performed oral sex on her 9.9 8.8 H e forced her to touch his genitals
through his clothes 2 9 0.7 He forced her to touch his genitals
under his clothes 5.8 2.2 He forced her to perform oral sex on him 2 5 4.4
He forced her to have sexual int emmrse 20.6 15.3
These are the rerultr of a survey of 380 women and 368 men enmlled in introductory psychology c w ~ e s st Teur h&M University. Percentages add up to more than 100 because often more thm one tinwanted wmal aaiviiy occurred on one date. Source: B a e d on Muehlenhard and Linton 1987:iW.
7. ANALTZINC THE RESULTS Mer the data are gathered, it is time to analyze them. Sociologists have spe- cific techniques for doing this, each of which requires special training. They range from statistical tests (of whlch there are many, each with i t s own rules far application) to content analysis (examining the content of something in order to identify its themes-in this case perhaps magazine articles and teleri- sion reports about rape, or even diaries kept by women who have been raped). If a hypothesis has been part of the research (and not all wid research h a hypotheses), it i s during this step that it is tested.
8. SHARING THE RESllLTS
Now it is time to wrap up the research. In this step you write a report that shares your f i n h g s with the scientific mnlnlunity. You relate your findings to
the literature, to show how they are connected to what has previously been discwvered. You carefully explain your research pmedures so others can repli- cate them-i.e., can repeat the study to test its findings. In this way science slowly builds, adding finding to finding.
NOW let us look in g r e ~ t e r detail at t11e fifth step to exaniine the research methods that sociologists use.
S i x Research Methods
Sociologists use six research tnethodp (also called research designs). These pro- cedures for gathering data are surveys, secondary analysis, documents, experi- ments, unobtrusive measures, and participant observation.
SURVEYS
Let us suppose that you want to know how many women are raped each year. The wmy-having people answer a series of question-would be an appro- priate method to use.
Before using this method, however, you have to decide whom you will survey. What is your populatim; that is, what i s the target group that you want to learn about? Is it a I I females in the world? Only U.S. or Canadian femdes? The females in a particular state, county, or city? Only females above a certain age? Or only those on your college campus?
Let us suppose that your research interest is modest-that you want only to know the extent of rape on your campus. Ideally, you would survey dl female students. But let us also s u p p s e that your college enrollment is large, mahng this impractical. To get at the answer. then, you must select a smaller group, a sampk of individuals, from whom you can generahe to the entire campus. Choosing a sample is critical, for it will a f f e ~ * the results of your study. For ex- ample, you d get different results if you survey only freshmen or seniors--or only women talang introductory sociology or advanced physics classes.
What kind of sample d allow you to g e n d i z e to the entire campus? The best is a r a h sample. This does nut mean that you stand on some cam- pus comer and ask questions of whomever happens to walk by. In a mn&m sample everyone in the pupubtwn has the same chance of being incluukd in the study. I n this case, since the population is ail women taking classes at your college, a11 such women must have the same chance of being included in your study-whether they are freshmen, sophomows, juniors, seniors, or graduate students. It also means that such factors (oariaBles) as a woman's choice of major, her grade point average, o r whether she is a day or evening student can- not affect her chances of being a part of your sample.
To obtain a random sample, you would need a list of aU women currently
How Sociologists D o Research / 37
enrolled in your college. TO determine wl~ich students become part of the sample, you might assign a number to each name on the list and then use ran- dom numbers to determine which particular persons become part of the sam- ple. (Random numbers are available nn tables in statistics books, and can be generated by computer.)
Because a random sample represents the population (in this case woinen students at your college), you can generalize your findings to all the women students on your campus, whether they were part of the sample or not.
I n some surveys, questionnaires, a list of questions, are mailed to people. Although such self-administered questionnaires allow a large number of people to be sampled at a lower cost, control is lost. For example, under what condi- tions did people (respondents) fill them out? Who influenced their answers?
Other surveys use intepuietlls: Respondents are asked questions directly. This is usudy done on a face-to-face basis, although some interviews are con- ducted over the telephone. The advantage of this type of survey is that the re- searchers bring control to the situation. They know the conditions under which the interview took place and that each question was asked in precisely the same way. Its &sadvantages include not only the more limited number of questionnaires that can be completed, and the increased cost, but also inter- viewer bias, the effects that interviewers can have on respondents that lead to biased answers. For example, although respondents may be d i n g to write an anonymous answer, they may not want to express their opinions to another person h e c t l y . Some even tly to make their answers match what they think the interviewer wants to hear.
Sociologists sometimes use closed-ended qttestwns, c d e d structurecl in- terviews. Each question is followed by a list of possible answers. The advan- tages are that these are faster to administer, and make it easier for the answers to be coded (categorized) so they can be fed into a computer for analysis. If you use closed-ended questions, you will have to be careful to make sure that they represent people's opinions. For example, if you were to a s k "What do you think should be done to rapists?" and the only choices you provide are to castrate or kill them, you would not be taking accurate measurements of peo- ple's opinions. Similarly, if you begin a question with, "Don't you agree that" ("rapists should be locked up for life"-r whatever you want to add), you would tilt the results toward agreement with a particular position.
Questions, then, must be carefully worded so they do not slant answers- because biased findings are worthless. It takes a great deal of training to con- struct questions that are free of bias, and sociologists are extremely critical of both how questions are worded and how they are administered (given).
To better tap the depth and diversity of people's experiences and atti- tudes, you may wish to use q e n - e n d e d questions, called ~lnstarcturerl inter- views, that allow people to answer in their own words. The primary advantage of this type of interview is that it allows people to express their full rang of opinions. The major disadvantage is that it is dfficult to compare people's an-
38 1 J a m M. H e d i n
fivers. For example, how would you compare these answers to the question '%'hat do yo11 ttiirlk causes rape?
'They haven't been raised tight." "I think they must have had problems with their mother." 'We ought to kiU every one!" "They're dl sick." "I don't want to talk about it."
The research topic we are considering also brings up anothe~ significnnt item. Let us suppose that you want to inteniav rape victims. Would they re- dly give honest answers? Will a woman even admit to a stranger that she has been raped, much less talk about i t ? Wouldn't all your efforts be Futile?
If you were to simply w a l k up to a stranger on the street and ask if she had ever been raped, you can guess the resultz-atd they certainly would give little basis for placing confidence in your findings. Researchers must establish mpp& (pronounced nth-pour), a feehg of t r u s t , with their respondents. When it comes to sensitive topics. arms about which ~ e o p l e may feel er~tbarrassment, shame, hostility, or other deep emotions, rapport is all the [nore important.
Once rapport is gained (often through b d h g twit hy erplailling the sig- nificance of the research. assuring anonymity, ant1 first ~ k i r i g nonsensitive questions), victims usually will talk a b u t rape. For example, each year re- searchers conduct a national crime survey in whch they ~nteniew a random sample oC 49,000 households-airnut 90.000 Americans. They find that most rape victims d talk about thelr expriences. The national crime surveys show that rape is twice as high as the official statistics, and that most rape is commit- ted by someone the victim knows (Schafran 1995).
SECONDARY ANhLf'SLS
In secondary annlysis, the second research method we shall consider, the r e - searcher ar~d).zes data already collected by others. For example, if you were to examine the bait data gathered by the interviewers who hd the national crime survey just mentioned, you would be doing s e c o n d q analysis.
Ordinarily, researchers prefer to gather their own data, but lack of re- sources. especially time and money, may make this preference impossible to fulfill. In dditian, data already gathered may contain a wealth of information not pertinent to the purposes of those who did the original study. It simply Iies there, waiting to be analyzed.
While this approach can solve problems of access, it &o poses its own ptoblems. Not having directly carried out the research, how can you be sure that the dilta were systematically gathered and accurately recorded, and biases avoided? That [nay not be an easy task, especially if the original data were gathered by numerous researchers, not all of whom were equally qualified.
How Sociologists Do Research 1 39
The use of h c r a m : n t s , written sources, is a third research method employed by soc~ologists. To ~nvestigatt. swial life, sociologists examine such diverse sollrces as books, newspapers, diaries, bark records, pnlice reports, immigra- tion rrcurds, household accoitnts, md records kept by various organizatio~ls. (Although t h e y are not commonly called documents, also included here are mrxles. television program$, videotapes, corrjputer disks, and CD-ROMs.)
Tu apply this mcthod to the study oi ripe, you might examine potice re- ports They may reveal what propti011 oT aU arrests are for rape; how many of the men arrested go to trial; what proportion is convicted, put on probation, sent to prison; hnd so forth. If these are your questions, police statistics could be valuable.
But for other questions, police records would be useless. For example, if you want to know about the adjustment of rape x ~ c t i r n s , they would tell you nothing. Other documents, however, may lend then~selves to this question. I f your campus has a rape crisis center, for example, tt m~ght have recards that would provide key information. Or you may obtain diaries kept by vichms, and search them for clues to their r e a c t i ~ n s ~ s ~ e c i a l l y how thew orientations change over time. If you couldn't find such diaries, you might contact a sample of rape victims and ask them to keep daries. h a t i n g that sample 1s extremely difficult-but, again, the rape crisis center could be dw key Their personnel might ask victims to keep the haries. (To my knowledge. h o w e v ~ r , no sociolo- gist has yet studied rape in this way.)
I am writing, of course, about an ideal case, as though the rape cris~s cen- ter is opening its arms to you. In actual fact it may not cooperate a t 41, refus- ing to ask victims to keep diaries and not even letting you near their retrords Access, then, is another problem researchers constantly face. Simply put, you can't study something unless you can gain access to it.
EXPERIMENTS
A Fourth research method is the experiment This is the classic method of the ~latural sciences. S ~ i o l o g i s t s seldom use it, however, because they are more likely to h e interested in broad features of satjety and social behavior. or in studying a social group i n a natural setting, neither of which lends itself to an experiment
The basic purpose of a n experiment is to identify cause-and-effect rela- tionships-to find out what caiises what. Ordinarily. experiments are used to test a h}potl~esis. Experimerlts involve i n d e p e n h t variables (those factors that cause a change in something) attd depencknt undubles (those factors that are changed). Before the experinlent, vou must measure the dependent vari- able. Then, after introducing the indepettdent variable. you again measure the rinpe~ldent variahle i n order to see what cllarlge has occurred.
40 1 James M. Zlawlin
Ldt us assutne, for example, that you want to test the hypothesis that pornogmphy creates attitudes that favor r a p . The independent variable willd be pornography, the dependent variable attitudes toward rape. You can mea- sure a group of men's attitudes toward rape and then us? random numbers to divide the men into two subgroups. To one group, the expwimental group, you introdu~e the independent variable (such as violent prnographic movies). The other group. the control group, is not exposed to the independent variable i t l l a t is, they are not showfi these movies). You then meas.1u.e the dependent variable i r b both groups. CIlanges in the dependent variable (in this case atti- tudes toward rape) are due to what only the experimental group received, the independent vuiable (in t h i s case, the pornography).
Because there is always some chance that unknown third variables have not been evenly distributed among the groups, you would need to retest your results by repeating the experiment with other groups of men.
UNOBTRUSIVE MEASURES
The fifth method we shdl consider is utwbtmsiw masrrra- o b s e ~ n g swial behatior when people do not h o w they are being studied. For example, smid researchers havc studied the 1eveI of whiskey consumption in a "d j' town by cwnting empty bottles in trash cans; the degree of fear induced through telling ghost stories by measuring the shrinking diameter of a circle of seated children; md the popularity of exhibits at Chicago's Museum of Science and Industry by the wear upon ties in front of the various displays (Webb et al. 19%).
Unobtrusive llieasures could also be used to study rape. For example, )nu could observe rapists in prison when they do not h o w they are being watched. You might arrange for the lea&r of a therapy group for rapists to be called out of the rwm. During his absence, you could use a one-way mirror to observe the men's interactions, and video cameru to preserve what they say and do. Such an approach would probably tell you more about their real attitudes than most other techniques.
Professimal ethics, however, might disallow such a shdy. And 1 know of no research that ha applied this methcd to the study of rape.
PARTICIPANT OBSERVATION (FIEIO WORK)
Let's turn to my favorite method, one that involves h e *searcher in the most direct way. In participant ohssrvation ( o r field work) the researcher partici- pates in a research setting while ohm'ng what is happening in that setting.
How is it possible to study rape by participant observation? I t would seem that this method would not apply. I f one considers being present during rape, i t certainly does not. But there we many other questions about rape that can be answered by participant observation, answers h a t cannot be gairied as ade- quately by any other rnetm.
How Swiologists Do Research / 4 1
k t us SUppOSr that your interest i s the adjustment of rape v i 6 t n s . Y o u wr,llld like to learn how the rape has al'fected their behavior and thcir orienta- tions to the mmrld. For example. how has their victimization affected their hopes and goals, their dating patterns, their ideas about men and intimacy? Participant abservatinn can provide detailed answers to such questions.
k t ' s go back to your clrnpus again. hsume that, like rnine, your campus has a rape crisis intervention center. This setting lends itself to participant ob- servation, for here you can observe ripe victi~rls fioin the time t h ~ y first report the attack ta their later prticipation in indi~idual and group counseling. With gmd rapport, you can even spend time with victims outside this setting, nh- serving how it affects other ;u;~wcia of their lives.
Participant observation has the ildded beneiit of albwing you to study whatever happer~s to orcur while you are it1 the setting. In this instance, you would also be able to study the operatio11 of the rape crisis center. A s you observe counselors at work, you could alstl study thcir attitudes and behaviors.
As you mav have noticed, in participant observation personal chara~-ens- tics of the researcher become highly important. Coulti a male researcher, for example, mnduct such research? TechnicalIy, the annver is yes. Properly in- troduced and with the right attitudes, male sociologists could do this research, But granted the topic, and especially the emotional states of females who have been brutally victimized by males, it may be more appropriate for female swi- ologists to conduct t h i s research. 'Their chances of s t ~ i c e s s are likely to be higher.
In conducting research, then, sociohgists must be aware of such variables as the sex, age, race, personaliiy, and wen height and weight of the researcher (Henslin 1990). \Vhile inlprtant in all research m e t h d s (for exampte, male respondents to a survey may be more talkative to young, shapety fernalr re- searchers than to obese males), these variables are especially i~npclrtant in par- ticipant obsenation (Snyder 1982).
Participmt observers fam a problem wid1 genedzability. Although they look for principles of human behavior, it is difficult to know the extent to which their findngs q p l y beyond the setting in which they occur. Consequently, most participant observatiorl is exploratory in nature: The find- ings document in detail what people in a particular setting are experiencing and how they are r e a d n g to those elrperiences. suggesting that other people who face similar situations w d react in similar ways.
I find participant observation the most exciting of the methods. It is the type of swiology that 1 like to do and the type I like to read about. From these studies, I gain a depth of understandmg of settings that I want to know more a b u t bur for whatever reason am not able to study, and in some caxes am not even able to enter. If I were a woman, for example, I might have volunteered for work in my campus's mpe crisis center-a technique often used by sociulu- gists to solve the problem of access.
Conclusion: A Note on Choosing Research Methods
As you have seen, a critical factor in choosing a research method is the ques- tions you wish to answer. Each m e t h d lends itself much better to answering particular interests or questions than do other m e t h d . You also have seen that access to subjects is c r i t i d in deciding which research method to use. TWO other factors me signrficant in this choice: the resouroes available to the researcher, and the researcher's background or training. For example, a re- searcher who prefers to conduct a survey may find that finances will not per- mit it, and instead turn to the study of dwuments. The researcher's hack- ground is similarly significant in t h i s choice. Researchers who have been trained in quantZtatim techniques (an emphasis on precise memurement, numbers, statistics) are more likely to use surveys, while researchers who have been trained in qualitative techniques (generally. making hrect observations of what p p l e do and say) lean toward participant obsewation. The particular training that sociologists reoeive in graduate school, which sometimes depends on capricious events, orients them tow& certain research methods. They feel comfortable with those, and tend to continue to use them throughout their ca- reers.
References
Henslin, James M. (1990). "It's not a lovely p k to visit, and I wouldn't want to Iive there." In Robert G . Burgess (ed.), Studies iin Quditaaloe Methodoiogy 11: 51-76. Greenwich, C T JAI Press,
hiuehlenhard, Charlene L., and Melolney A. Ljnton (1987). "Date rape: Familiar strangers."]ou+nal~Cmnsellng t'sy&k~gy 34: 186-96.
Schaefer, Richard T. (1989). S&gy. 3d cd. New York: McGraw-HiU. Schafran, Lrnn Hecht. 'Rap Ir Still Underreprtd." New Ymk Times, August 29,
19!#:19. Snyder, Mvlarb; (1982). *Self-fulfilling stemtypes." Psychology Tothy, July: 60, 65,
6749. Webb, Eugene J.. Donald T. Campbell, Richard D. Schwa*, and b e Sechrest (1966).
Unobtmstw ,U~anrres Nonreoctiue &seam11 in the Sodal Sciences. Chicago: Rand McN ally.