|
William Griffitt and Elaine Hatfield METHODS AND ISSUES IN SEX RESEARCH Excerpt from the textbook: Griffitt and Hatfield: Human Sexual Behavior, pp. 6-13, Scott, Foresman & Co., Glenview (IL) and London, 1985. Reproduced here by permission of the copyright holders. 1. Evaluating Sex Research A substantial amount of what we know (or think we know) about sexual behavior and sexuality has been learned through the scientific research efforts of the twentieth century. Most people who are not directly involved in it, however, find science and scientific research a bit puzzling: For most people, "science" evokes images of men in white jackets surrounded by gleaming equipment or memories of biology labs in which the "experiments" never quite worked out the way the lab manual said they should. To some, whatever scientists do seems incomprehensible and somewhat frightening, like the activities of a powerful secret society. (Byrne, 1974, p. 3). But scientific research is not really very mysterious. Scientific research in sexual behavior simply involves the use of a few procedures by which scientists attempt to obtain and evaluate information with the goals of improving our abilities to understand, predict, and possibly exert some degree of control over sexual behavior. How can we judge how adequate their attempts are? 1. Evaluating Sex Research Sampling. In most sex research the goals are to achieve accurate descriptions of the sexual behaviors of a population and to describe the relations between these behaviors and other attributes of the members of that population such as, for example, personality characteristics or religious background. Another frequent goal is to learn the relations between sexual behaviors and the events or objects to which people are exposed (for instance, genital stimulation, pornography, people of the other or same sex). In several ways the accuracy of the descriptions obtained and conclusions reached depends on who is studied. In most research it is impossible to study all the people in the population of interest (for example, all American men and women, all married couples over sixty-five, all homosexuals). Instead, scientists must study smaller samples of people and attempt to generalize the sample findings to the larger population. For example, many investigators have sought to estimate the percentages of American men and women who engage in heterosexual intercourse before marriage by studying only limited numbers of men and women from the American population. It should be apparent that in such investigations, who is studied—the composition of the samples—has an important bearing on what is found and on the accuracy of the generalizations that may be made regarding the larger population. A representative sample of adequate size is needed for accurate generalization. Representativeness is the degree of match between the sample and the larger population in such characteristics as religion, liberalness, educational level, and other important variables. For instance, if the sample contains more liberals than does the population as a whole and if liberals tend to have more premarital intercourse than conservatives, population estimates of the prevalence of premarital intercourse will be too high. Sample size is important because the accuracy of statistical estimates tends to increase as the sample size increases. An enormous sample size cannot "make up" for the limited accuracy of a nonrepresentative sample. A sample that is too small for accurate estimates can, however, seriously weaken a study with a representative sample. Even if a large and representative sample is initially selected, another problem remains. Because of the sensitive nature of the topics of interest, some people refuse to participate in sex research. For example, in the Playboy Foundation-sponsored survey (Hunt, 1974) only 20 percent of those people asked to participate actually did so. The problem is that volunteers for sex research may be very different from those who refuse to participate, and the conclusions reached from volunteer data may therefore be inaccurate. Research Procedures. In judging the extent to which findings may be taken as applicable to people other than those sampled, it is an important to know from whom sex information is obtained. Also important is how the information is obtained. There are many possible ways to obtain sex information about people, and each has its advantages and disadvantages. One approach is simply to ask people to report on their sexual behavior through interviews or questionnaires. A major issue here is the extent to which self-reports are accurate reflections of people's actual sexual behavior. Respondents may underreport, overreport, or do both when responding to sex interviews or questionnaires. Underreporting may occur for at least two reasons. First, people may have engaged in certain sexual activities that they wish to conceal because of guilt, fear of censure, or fear of disrupting their marriages, social standing, or careers. This is most likely when information concerning taboo topics such as homosexual or extramarital heterosexual involvements is requested. For this reason it is important that respondents be assured of the privacy and anonymity of their responses. Second, people's memories may fail them. They may simply forget about sexual experiences or be unable to recall accurately how often they engaged in certain activities at certain times in their lives. Overreporting may occur when people seek to bolster their sexual image by claiming experiences they have never had. They may be embarrassed because they lack sexual experience and simply expand a sexual history to more fully coincide with what they think is "normal" or desirable. Both under- and overreporting may occur when, for example, a person conceals homosexual involvement and exaggerates heterosexual accomplishments. Such problems may be minimized when investigators communicate to respondents full assurance that all experiences are equally valued and that honesty is important. Furthermore, questionnaires and interviews can be and often are designed with built-in checks for accuracy such as the repetition of questions. Information may also be obtained through direct observations of behavior as in Masters and Johnson's (1966) investigations of physiological responses to sexual stimulation or the observational work of Hooker (1965) in the gay community. Although observational studies are more difficult and time consuming than self-report studies, they minimize some of the problems of obtaining data distorted by respondents' self-interests. Still, the mere presence of observers may influence the behavior of respondents in unknown ways. Are peoples' sexual responses the same when they are knowingly being watched as when they are in the privacy of their bedrooms? Do homosexuals behave the same when they are being observed by a "straight" investigator as when they are alone? A second issue in observational studies concerns the ability of observers to accurately record what they observe. The use of sophisticated recording instruments can in some cases minimize inaccuracies, but such equipment is often not accessible or appropriate. Most often observers must either commit observations to a possibly faulty memory for later recording or unobtrusively take notes at the time the behavior is taking place. Observer bias, another problem, occurs when the observers' own attitudes influence their interpretations and reports of observed events. Others may simply not "see" the same thing. Though these are only some of the methodological issues involved in sex research, they are important considerations in assessing the quality of research and interpreting research findings. We hope that readers adopt the habit of considering such issues in their evaluations not only of the research discussed in this book but also of the tremendous amount of sex "information" and advice found in almost every issue of daily newspapers and popular magazines. 2. Methods of Sex Research Case Study Methods. Much early sex research involved intensive in-depth studies of the lives of individuals or small groups of individuals. In one common type of case study, for instance, an individual's life is explored extensively. From the quantities of information thus collected, the investigator can learn much about the sexual preferences, experiences, conflicts, desires, and aversions of that individual. In addition, the patterns of meanings associated with the individual's sexuality and the circumstances surrounding his or her sexual development can be examined in great detail. Usually such studies are unstructured and open-ended with respect to the topics covered, and far more information about the individual is obtained than in survey or experimental investigations. (Online editor’s note: Famous historical examples are the case studies of Sigmund Freud. EJH.), Survey Methods. Though case study methods obtain a great deal of information about a small number of people, survey methods typically acquire a relatively small amount of information from a large number of people. Because of this, there is the potential for both the representativeness and the size of survey samples to be more adequate for generalization to larger groups of people. As we noted earlier, however, large sample sizes provide no guarantees of representativeness. For example, even though the large-scale questionnaire studies conducted by Redbook magazine (Tavris, 1978a, 1978b; Tavris & Sadd, 1977) obtained responses from over 100,000 women and 40,000 men, the samples were underrepresented by those over the age of fifty, those with limited educations, and nonwhites. Similarly, the widely acclaimed and monumental studies of Kinsey and his colleagues (1948, 1953) involved interviews with a sample of nearly 12,000 people that, for reasons to be discussed later, could not be considered representative of the population of American men and women. A second potential advantage of survey methods is that the procedures used for gathering data are usually fairly well standardized and the method of questioning is more or less consistent from one respondent to another. That is, there may well be less potential for error in recording responses than there is in case studies. Whatever their shortcomings, survey methods can provide quite useful information concerning the prevalence of certain sexual activities, sexual attitudes, and the relationship among these variables and other characteristics of respondents. Experimental Methods. One of the important shortcomings of both case study and survey methods is that it is virtually impossible to make causal interpretations of observed relationships between or among variables. For example, in a study of male and female college students, Griffitt (1975) found that those with a great deal of sexual experience were more erotically responsive to photographs of explicit sexual activities than were those with limited sexual experience. How might this finding be interpreted? At least three possibilities exist. First, it might be that as a result of increasing degrees of sexual experience people become more sexually responsive. Second, the reverse could be true, people who are most sexually responsive might be more inclined to become sexually active. Still another possibility is that some third factor such as sexual liberalness accounts for the relationship between experience and responsiveness. Those who are most sexually liberal acquire more sexual experience and are more responsive to erotic stimuli than are the sexually conservative. Without some method of identifying which of these three variables (experience, responsiveness, or liberalness) controls the other two and any additional potentially important variables, there is no way to determine cause and effect relationships among them. The primary advantage of experimental methods is that they involve procedures in which substantial control may be exerted over the critical variables of interest allowing for the possibility of drawing causal inferences. For example, in their experimental investigations of male and female sexual response cycles, Masters and Johnson (1966) were able to demonstrate the effects of orgasm on the physiological release of sexual tension. When women who were highly sexually aroused were allowed to reach orgasm, the many physiological "symptoms" of sexual arousal, such as tissue engorgement with blood and muscular tension, dissipated much more rapidly than when the women were not allowed to reach climax. Thus, orgasm was shown to play an important causal role in the resolution of sexual tension. A disadvantage of many experimental studies, however, is that they obtain data using relatively small samples of respondents whose behavior may or may not be generalized to larger populations. For example, Masters and Johnson (1966) studied comparatively few people and only those who were able to respond and perform sexually while attached to various recording instruments and being observed and filmed. The extent to which the findings may be generalized accurately to people who would not be "caught dead" in such circumstances is open to debate. Similarly, much experimental sex research is conducted using college freshmen and sophomores as subjects, and it is unclear how accurately investigators may generalize from such samples to the larger population of interest. These, then, are the major methods used in sex research. Most other methods discussed in this book (for instance, cross-cultural and participant observation) are variations of the case study, survey, or experimental approaches. Each approach has its advantages and disadvantages, and ideally, of course, all three would be used in any investigation of a particular form of sexual behavior (for example, masturbation) to capitalize on the unique merits of each. 3. Ethical Issues in Sex Research Informed Consent. The principle of informed consent requires that research subjects be fully informed of the potential risks and harm associated with participating in the research, the general type of information to be obtained in the research, and the manner in which their own information will be used. They must understand that they are free not to participate and that they may withdraw at any time even if they do agree to participate. When minors are involved, their parents or guardians must provide informed consent. Protection from Harm. According to the ethical principle of protection from harm, participants must suffer no harm from the research, or the likelihood and degree of possible harm must be outweighed by the benefits of the research to the participants or the larger society. How might participants in sex research be harmed? First, some of the behaviors (for example, homosexual acts, oral-genital contacts, and extramarital intercourse) are illegal in many states or at least considered socially undesirable by many people. For this reason, the anonymity of research participants must be carefully guarded. Leakage of the fact that a person has engaged in any of these acts may lead to legal prosecution, marital disruption, professional ruin, or other negative consequences. Second, most sex research focuses on topics that some people may find emotionally distressing. For example, even though childhood homosexual experiences are rather common, adults who are asked to recall and discuss such experiences may react with guilt, shame, or embarrassment. For this reason, researchers must attempt to use procedures that minimize such stressful reactions in any way that is feasible. These, then, are some of the major methods and issues involved in sex research. For more detailed discussions of methodological issues, interested readers should refer to the work of Bentler and Abramson (1981). A thorough examination of ethical issues is provided by Masters, Johnson, and Kolodny (1977). Let us now turn to a brief overview of some of the major studies of human sexual behavior. 4. Modern Sex Research Survey Research The Kinsey Reports (1948, 1953). The two most widely cited and highly respected surveys of human sexual behavior were conducted and reported by Alfred Kinsey and his associates at the Indiana University Institute for Sex Research. The two volumes resulting from the surveys, Sexual Behavior in the Human Male and Sexual Behavior in the Human Female, described the sexual practices of nearly 12,000 American males and females. The information was obtained by use of interviews. The samples for the two studies were 5300 white men and 5940 white women. Those studied ranged in age from two to ninety and were drawn from all forty-eight states in existence at the time, from all educational and occupational levels, from several religions, and from rural and urban areas. In spite of the variety and rather large number of people in the samples, they were not representative of the general American male and female populations. For example, no blacks were in either of the samples. In addition, the well educated, college students, Protestants, urban residents, residents of Indiana and the Northeast, and young people were overrepresented in the samples. Underrepresented were laborers, rural residents and those from the West, less well-educated people, the elderly, Catholics, and Jews. Kinsey and his colleagues were fully aware that their samples were not representative. They were, however, convinced that no sample truly representative of the American population could be obtained for a sex survey. They were probably correct in assuming that refusal rates would be so high that attempts at random sampling would be futile. To compensate for weaknesses in the male sample, statistical adjustments were made in an effort to improve the accuracy of generalizations to the general population of American males. Nevertheless, a study committee of the American Statistical Association was highly critical of the Kinsey samples (Cochran, Mosteller, & Tukey, 1954). Although weaknesses in the Kinsey samples make generalizations to the population risky, the research procedures used with the obtained samples are praiseworthy. The basic interview consisted of 350 questions and required about ninety minutes to conduct. The interviewers memorized all the questions and recorded respondents' answers in a code known only to themselves. Moreover, they developed techniques for putting respondents at ease and keeping them at ease as sensitive topics were raised. Their approach was nonjudgmental; all reported sexual experiences were recorded with acceptance. Questions were asked directly and the interviewers were adept at and free to adjust their language to that favored or most easily understood by the respondents (Pomeroy, 1972). All of these and the many other interviewing procedures employed were designed to obtain as accurate information as possible. Also included were built-in checks and crosschecks to detect fabrications. The data obtained by the various interviewers were compared. Some respondents were interviewed twice, and the answers of wives and husbands were often compared for agreement. Thus, it was determined that incidence data (responses that indicated whether or not respondents had ever engaged in certain behaviors) and vital statistics data (age, occupation, and so on) were fairly accurate. On the other hand, frequency data (how many times people had engaged in sexual acts) were less accurate, perhaps because respondents' memories failed them. In short, even though the sampling procedures were relatively poor, the excellent interview procedures seem to have produced fairly accurate information from those actually studied (Cochran, Mosteller, & Tukey, 1954). The Kinsey Reports stand out as monumental pieces of work in the field of sex research. Morton Hunt (1974). In 1972, Research Guild, Inc., a Chicago-based independent market-survey and behavioral research organization, surveyed American sexual behavior under the commission of the Playboy Foundation. The data were analyzed and presented by Morton Hunt, a professional writer, in Sexual Behavior in the 1970s (Hunt, 1974). Most of the data were obtained with written questionnaires but were supplemented by interviews with an additional 100 men and 100 women. The final sample consisted of 982 males and 1044 females. These respondents were selected from the phone books of twenty-four cities around the country. Potential respondents were contacted by phone and asked to take part in small groups to discuss sexual behavior. No mention of a questionnaire to be completed was made. About 20 percent of those contacted agreed to participate (an 80 percent refusal rate). Because names were selected from the phone book, young people were underrepresented in the original sample; to correct this, several hundred young people were added by individual solicitations. But again because of the selection method, several other groups such as prison inmates, illiterates, and the very poor were either underrepresented or passed over entirely. According to Hunt, the final sample closely matched the general population eighteen years old and over in age, occupational status, education, urban-rural background, and attitudes on a number of different issues. No empirical evidence for these claims is presented, however. Even if it were, the fact that 80 percent of the initially randomly selected sample refused to participate raises serious questions about how representative the sample actually was. Were those who refused less sexually experienced, less open about sex, less interested in sex, or more easily embarrassed by sex than the participants? We do not know. Hunt acknowledged that the sample was not perfectly representative of the American adult society. But his belief that it was "reasonably good" (p. 16) led him to draw inferences that were frequently phrased as if they were accurate generalizations to the population as a whole. Two procedures were used to obtain information. The actual data were collected using four different versions of a questionnaire. The one completed by a given respondent depended on the participant's sex and marital status. The questionnaires contained between 1000 and 1200 questions. Even if some could be skipped, this is a formidable number of questions to ask. Boredom, fatigue, lack of attention, and other factors may have resulted in many inaccurate replies. The questionnaire data were supplemented by interviews with 100 men and 100 women. The Hite Reports (1976, 1981). In 1976 Shere Hite published The Hite Report, an analysis of female sexuality based on the responses of women to questionnaires prepared by the author. Partly because of its controversial conclusions and a tremendous amount of media promotion, the book received wide public attention. In 1981 the same author published a companion book titled The Hite Report on Male Sexuality. Both books have been highly controversial, drawing praise from some but severe criticism from others on sampling and procedural grounds. The female report recorded and presented some summary statistics of the responses of about 3000 women who returned questionnaires sent to over 100,000 women. The women who returned questionnaires were from many parts of the country, ranged in age from fourteen to seventy-eight, and represented a variety of religious, educational, and professional backgrounds. In spite of this, however, the sample was definitely not representative of the general population of American women. The 97 percent refusal rate virtually guarantees this. In addition, the questionnaires were distributed with assistance from many women's groups (such as the National Organization for Women) with decidedly feminist memberships and from publications such as Ms. and the Village Voice. Oui magazine published the entire questionnaire. Because of the low response rate and the manner in which questionnaires were distributed, it is likely that the final sample of women was considerably more sexually liberal than the general population and perhaps different from it in many other ways as well. For the male report, 119,000 questionnaires were distributed and 7239 returned—a 6 percent return rate. The questionnaire distribution method was similar to the earlier study and included reprinting the questions in such magazines as Penthouse. Even though its exact characteristics are difficult to determine from the book, the final sample is not representative of American men. Among other differences, the sample appears to be more sexually liberal and more economically and educationally advantaged than the general population (Robinson, 1981). In short, the sampling methods for both reports leave much to be desired if one is seeking representativeness. The author, however, did not attempt to generalize her findings to the general population. They are presented mainly as insights into the sexuality of some American women and men. Both reports rely mostly on open-ended questions in which respondents were asked to elaborate on their experience with and feelings about various sexual activities. The female questionnaires included about 60 such questions and the male questionnaires 183. Many of the questions required lengthy replies, and respondents were encouraged to omit questions if they wished. Some questions were rather bluntly and, for some respondents, possibly offensively phrased ("How do you masturbate?" "Do you like rectal penetration?") or even leading ("Are most of your partners sensitive to the stimulation you want?"). Many questions undoubtedly went unanswered; one must wonder how those who answered certain questions differed from those who did not. Though they cannot be regarded as accurate barometers of the sexuality of American females and males, the detailed statements of the respondents are a rich supply of information about the sexual concerns of at least some people. Laboratory Studies Using Direct Observations Though it was not widely publicized, the Kinsey research team used direct observations of sexual activity in compiling knowledge about the physiological aspects of sexual responses (Pomeroy, 1972). Their preliminary findings were reported in the 1953 volume dealing with female sexual behavior. But it was the research team of William Masters and Virginia Johnson that first brought to national and worldwide attention the use of direct observation of sexual activity in the laboratory as a major research method. Masters and Johnson (1966, 1979). In Human Sexual Response (1966), Masters and Johnson summarized several years of research concerning the physiological changes that take place in humans in response to sexual stimulation and during sexual activity. …Though their initial reports (1966) concerned physiological responses in heterosexuals during masturbation and heterosexual activity, they later confirmed that responses during homosexual activity and in homosexuals are virtually identical. Their findings for homosexuals are reported in Homosexuality in Perspective (1979). The 1966 report is based upon observations of 382 women and 312 men as they engaged in sexual activity in a laboratory setting. The sample is not representative of the general population of American women and men. The women ranged in age from eighteen to seventy-eight and the men from twenty-one to eighty-nine. The final sample consisted of 276 married couples, 108 unmarried women, and 36 unmarried men. Most participants were associated with the Washington University of St. Louis community and were selected only if their sexual anatomy was normal, if they could respond sexually in a laboratory setting, and if they were able to describe in detail their sexual reactions. They were mostly white, well-educated urban dwellers of above average socioeconomic level. Most important, however, they were willing to perform sexually while being observed, photographed, and attached to physiological recording equipment. Even though the sample was not a representative one, there is little reason to suspect that the recorded physiological reactions of the respondents differ dramatically from those of other people. The second observational study included a sample of 94 male and 82 female homosexuals who were compared with two heterosexual groups. The first heterosexual group consisted of 57 male and 57 female heterosexuals selected specifically for the new study. The second group was formed from 286 male and 281 female heterosexuals who took part in the original study. The homosexual participants were recruited from various parts of the country but it is unlikely that they were a representative sample of American homosexuals. Their educational level was higher than average, they were more likely to have a committed stable relationship with one person than most gays, their ages were concentrated in the twenty to forty range, and they were specifically selected for a history of easily reaching orgasm in a variety of ways. Nearly all were white. Furthermore, there are many difficulties involved in designating persons as heterosexual or homosexual Some critics have argued that Masters and Johnson's selection of their homosexual participants was seriously flawed by their failure to adequately consider such problems (Zilbergeld & Evans, 1980). Many of the same procedures were used to gather data in both of the studies. Participants were interviewed concerning their sexual histories and given medical exams. As in the original study, participants were observed in a variety of sexual activities. These included masturbation, manual stimulation by partners, fellatio, cunnilingus, coitus, and anal intercourse. Visual observations, film, and recording instruments were used to record body changes during sexual activity. Cardiovascular reactions, muscle reactions, skin reactions, and respiratory reactions were recorded. A special artificial plastic penis that contained a light source and photographic equipment was used to record vaginal, cervical, and uterine changes during sexual responses. Participants were interviewed after their sexual activities. The Masters and Johnson work is important but not without its flaws. Many of their most important conclusions and interpretations are based not on their physiological data but on rather vaguely supported impressions gathered from the interview material. Their statistical breakdowns are confusing and difficult to interpret and other researchers would benefit from more complete descriptions of their study participants, particularly those participants who received treatment for their sexual problems (Zilbergeld & Evans, 1980). Masters and Johnson were the first to use direct physiological recordings on a large scale in the study of sexual reactions. Now, however, the use of sophisticated instruments to record sexual responses is rather common.
Bentler, P.M., & Abramson, P.R. (1981). The science of sex research: Some methodological considerations. Archives of Sexual Behavior, 10, 225-241.
Byrne, D. (1974). An Introduction to Personality. Englewood Cliffs, N.J.: Prentice-Hall.
Cochran, W.G., Mosteller, F., & Tukey, J. (1954). Statistical problems of the Kinsey report on sexual behavior in the human male. Washington, D.C.: American Statistical Association.
Griffitt, W. (1975). Sexual experience and sexual responsiveness: Sex differences. Archives of Sexual Behavior, 4, 529-540.
Hite, S. (1976). The Hite report. New York: Dell.
Hite, S. (1981). The Hite report on male sexuality. New York: Alfred A. Knopf.
Hooker, E. (1965). An empirical study of some relations between sexual patterns and gender identity in male homosexuals. In J. Money (Ed.), Sex research: New developments. New York: Holt, Rinehart, and Winston.
Hunt, M. (1974). Sexual behavior in the 1970’s. Chicago: Playboy Press.
Kinsey, A., Pomeroy, W., & Martin, C. (1948). Sexual behavior in the human male. Philadelphia: Saunders.
Kinsey, A., Pomeroy, W., Martin, C., & Gebhard, P. (1953). Sexual behavior in the human female. Philadelphia: Saunders.
Masters, W.H., & Johnson, V.E. (1966). Human sexual response. Boston: Little, Brown.
Masters, W.H., Johnson, V.E., & Kolodny, R.C. (Eds.). (1977). Ethical issues in sex therapy and research. Boston: Little, Brown.
Masters, W.H., & Johnson, V.E. (1979). Homosexuality in perspective. Boston: Little, Brown.
Pomeroy, W.B. (1972). Dr. Kinsey and the Institute for Sex Research. New York: Harper & Rowe.
Robinson, P. (1981, July). What liberated males do. Psychology Today, pp. 81-85.
Tavris, C., & Sadd, S. (1977). The Redbook report on female sexuality. New York: Delacorte Press.
Tavris, C. (1978, February). The Redbook report on male sexuality. Redbook, (a).
Tavris, C. (1978, March). The Redbook report on male sexuality. Redbook, (b).
Zilbergeld, B., & Evans, M. (1980, August). The inadequacy of Masters and Johnson. Psychology Today.
|