PART 4: CHAPTER 3

Social and Behavioral Science Research Analysis

Introduction

The Commission has examined social and behavioral science research in recognition of the role it plays in determining legal standards and social policy. This role, while notable, is not, nor should it be, the sole basis for developing standards or policy. The lack of funding and the inability under the mandate of the Charter to conduct original research has resulted in the need to rely on existing information. The amount of research conducted in the last fifteen years provides a reasonably sufficient base to reevaluate answers to old questions. Some might argue that given the controversy and heated debate that inevitably surrounds any discussion about pornography, in some ways, we might be better off relying on studies initiated, funded, and presented outside the context of such a milieu.

The major question which frames this research review is: what are the effects of exposure to pornography and under what conditions and in what kinds of individuals are these effects manifested? We also have structured this review with the following considerations in mind: (1) that it provides some input into the policy-making process; (2) that it provides social science information for public consumption and understanding; and (3) that it provides the research community with further questions for investigation.

While the nature of effects is the focus of this section, we have also examined public opinion on pornography to systematically describe the nature of public perceptions of and experiences with such material as well as policy preferences. In terms of effects, correlational as well as experimental studies on sexual offenders as well as on nonoffender populations were examined. For background purposes, we have also presented brief summaries of what some predecessor Commissions have concluded about the social science evidence before them.

Some observations on terminology and on the character of social science evidence are appropriate at this point as guidelines to reading through the rest of this chapter.

We will simply avoid the usual definitional morass by using the term "pornography" to refer to the range of sexually explicit materials used in the various studies reviewed here. In a number of studies, these materials have included sex education materials. In describing specific studies, we also will use the researcher's terminology of choice, but making sure that the stimulus materials are adequately described for the reader.

We also are sensitive to the limitations and strengths of specific research approaches and we have taken special efforts to review these briefly in each major section of this Chapter, if only to underscore the fact that our evaluation of the research recognizes these limitations and indeed proceeds from the assumption that any conclusions must be drawn on the basis of complementary or convergent data.

Overview of the 1970 Commission Research Conclusions

The period prior to the creation of the 1970 Commission on Obscenity and Pornography was marked by a paucity of research on the effects of exposure to pornography (Cairns, Paul and Wishner, 1962). A Commission-sponsored review of the literature in 1970 later concluded that "we still have precious little information from studies of humans on the questions of primary import to the law ... the data stop short of the 'critical point'". (Cairns, et al, 1970). Much of the Commission-sponsored studies thus constituted some of the earliest investigations on the issue of pornography.

The 1970 Commission funded over eighty studies to examine various aspects of pornography. Surveys included a national in-person survey of public attitudes toward and experiences with pornography (Abelson, et. al., 1970). A number of correlational studies examined social indicators of crime rates (Thornberry and Silverman, 1970; Kupperstein and Wilson, 1970; Ben-Veniste, 1970) while another cluster of studies investigated sex offenders and their previous experiences with erotica, patterns of exposure and self-reported arousal. Finally, another group of studies was commissioned (laboratory experiments) to examine causal links between exposure to pornography and effects (see Technical Reports of the Commission on Obscenity and Pornography, vols. 1, 6, 7, and 8, 1970).

The national survey findings (Abelson, et. al., 1970) showed that between two-fifths to three-fifths of the respondents believed then that sexually-explicit materials provided information about sex, were a form of entertainment, led to moral breakdown, improved sexual relationships of married couples, led people to commit rape, produced boredom with sexually-explicit materials, encouraged innovation in marital sexual technique and led people to lose respect for women (see comparison between 1970 survey findings and 1985 Gallup poll results below).

Experimental findings showed brief increases in sexual activities and fantasies after exposure to sexually-explicit materials but no significant alterations of established sexual behavioral patterns. The Commission further determined that there was no detectable relationship between availability of pornography and crime rates in the United States but suggested that removal of restrictions on pornographic material was correlated with lower sexual crime rates, as determined from Danish data prior to and after the removal of restrictions on pornography (Ben-Veniste, 1970; Kutchinsky, 1970, 1973).

The 1970 Commission concluded:

... In sum, empirical research designed to clarify the question has found no evidence to date that exposure to explicit sexual materials plays a significant role in the causation of delinquent or criminal behavior among youth or adults. The Commission cannot conclude that exposure to erotic materials is a factor in the causation of sex crimes or sex delinquency (p. 223).

The Commission's conclusions were challenged and a number of methodological issues were raised (Cline, 1974; Eysenck and Nias, 1980). At the very least, these conclusions were described as "premature" (see Liebert, 1976). Researchers who have done studies subsequent to the 1970 Report have also consistently identified a major flaw in the 1970 studies: the absence of any investigation of the effects of violent pornography.

On balance, however, the impetus for further research on the effects of exposure provided by the 1970 Commission cannot be overlooked. As the Effects Panel noted in its report,

One of the contributions of the work of the Panel has been to place the dimensions of human sexual behavior on the agenda for continuing inquiry. By providing resources in terms of funds and technical guidelines, the Panel has helped to legitimate systematic inquiry into an area that heretofore has either been ignored or feared.

It is difficult to quarrel with this observation.

Since the 1970 Commission report, in fact, numerous research studies have been done exploring various aspects of the effects of pornography. Since 1970 the quantity and quality of the research has been impressive. While much remains to be explored, not only has the volume of studies conducted steadily increased, but the programmatic nature of the research conducted by various individuals and research teams has provided a better insight into understanding the various conditions under which certain effects may or may not occur.

Studies done for the 1970 Commission were hampered by time constraints. As the research director for the 1970 Commission pointed out, "most of the researchers had less than nine months in which to establish a research team, arrange a research setting, develop measuring instruments, secure subjects, collect the data, reduce the data, and write a report." (General Preface to Technical Reports, Commission on Obscenity and Pornography, 1970, p. vii)

Methodological advances in measurement procedures have also enhanced the reliability and validity of research instruments and findings. For example, measures of sexual arousal in some of the 1970 studies were based almost entirely on self-reports (e.g., Cook and Fosen, 1970; Goldstein, et. al., 1970; Davis and Braught, 1970). Since then, the poor correlation between self-reports of sexual arousal to sexually explicit stimuli and physiological measures of arousal has been well documented (Abel, Barlow, Blanchard and Guild, 1977; Blader and Marshall, 1984).

More recent studies have used instruments such as the penile plethysmograph (Malamuth and Check, 1980a), thermography procedures (e.g., Abramson, et. al., 1981) or the vaginal photoplethysmograph (see Sintchack and Geer, 1975; Hatch, 1979) to evaluate arousal (see also Geer, 1975; Heiman, 1977), or have combined physiological measures (e.g., blood pressure readings) with paper-and-pencil tests. Researchers have also attempted to validate paper-and-pencil measures, a critical methodological requirement (see, for example, Burt, 1980; Malamuth, In Press). Finally, more sophisticated statistical techniques have allowed for better data analysis, control, and interpretation. Multiple regression techniques, for instance, have allowed researchers to specify how much each explanatory variable contributes to changes in the variable being measured. Various other statistical techniques have also helped in deciding whether correlational data give any credence at all to the possibility of causal linkages.

A final observation might be made with regard to stimulus differences between the 1970 studies and more recent ones. Stimulus materials used in the 1970 studies were obtained primarily from sex research institutes (the Institutes of Sex Research at Hamburg University in West Germany and at Indiana University) and the Bureau of Customs confiscated contraband collection. One researcher (Tannenbaum, 1970) resorted to producing his own film which he described as showing a young lady "going through the motions of disrobing in a fairly sensuous manner in apparent preparation for the arrival of a lover." These materials were also presented primarily in the form of slides, magazine pictorials, mimeographed passages and film.

It is perhaps as much a function of availability and changing technology that more recent studies have used as stimulus materials films, audiotapes, videos, and material from various "adult men's" magazines, all easily available from outlets as diverse as the neighborhood video store, the corner newsstand, or the local adult bookstore.

Other Pornography Commissions and Social Science Research

Other organizations which have studied pornography such as the Williams Committee in England and the Fraser Commission on Pornography and Prostitution in Canada have also examined social science research evidence on the effects of viewing pornography. (Report of the Committee on Obscenity and Film Censorship, 1979; Report of the Special Committee on Pornography and Prostitution, 1985).

The Williams Committee, working between 1977 and 1979, commissioned two reviews of the existing literature. One review examined the effects of viewing pornography (Yaffe and Nelson, 1979) and the other examined the effects of exposure to media violence (Brody, 1977). Both reviews highlighted the difficulties of studying human behavior and of understanding human motivations. The review of the effects of viewing sexually-explicit materials concluded that "there is no consensus of opinion by the general public, or by professional workers in the area of human conduct, about the probable effects of sexual material." The review on the effect of exposure to media violence similarly maintained that "social research has not been able unambiguously to offer any firm assurance that the mass media in general, and films and television in particular, either exercise a socially harmful effect, or that they do not."

The long track record of media violence research and antisocial behavior makes the latter conclusion somewhat surprising, particularly since an opposite conclusion was arrived at by a similar commission working under the direction of the United States Surgeon General in 1972, which had examined the effects of exposure to media violence (Surgeon General's Scientific Advisory Committee on Television and Social Behavior, 1972).

The conclusions of the Williams Committee on the effects of viewing pornography may not be as surprising since much of the experimental work was published after 1978. It is not clear, however, how much value these studies would have had for the Williams Committee since its call for more research was predicated on the importance of studying "the human personality as a whole, rather than to specific questions about violent or sexual materials and their supposed effects" (p. 4). The Committee further appeared to give greater attention to correlational studies as it examined in considerable detail studies by Court (1977) and Kutchinsky (1973). The Committee was highly critical of Court's methodology but also pointed out that the Danish data did not lead to the conclusion that the availability of pornography resulted in a decrease in sexual offenses.

The Canadian Fraser Commission similarly sponsored a research review (McKay and Dolff, 1985) and concluded that "the research is so inadequate and chaotic that no consistent body of information has been established. We know very well that individual studies demonstrate harmful or positive results from the use of pornography. However, overall, the results of the research are contradictory or inconclusive." (Report of the Special Committee on Pornography and Prostitution, v. 1, p. 99).

The commissioned review was exceedingly critical of the research, maintaining that the studies in every aspect exhibited "conceptually cloudy thinking," that they were characterized by "blatant silliness" and had no integrating framework, that "the literature is rife with speculation and unwarranted assumptions." The low regard for behavioral science methods is evident throughout the review, with major criticisms focusing on the uselessness of the experimental paradigm (p. 86-87), and the inability to draw conclusions from correlational research. Despite this assessment, the Commission proceeded to recommend criminal sanctions for sexually violent material and child pornography and limits on public display for nonviolent pornography. These recommendations were based on the Commission's observations that these materials were contrary to Canadian values of equality and human dignity.

It is obvious that the contribution of social science findings to policy considerations can vary, from being the sole or primary basis for policy recommendations, as was the case for the 1970 Commission, to being close to irrelevant to such considerations, as seemed to be the case with the Canadian pornography commission.

Public Attitudes Toward Pornography

How does the public view pornography and have there been any changes in public opinion in the last fifteen years?

Survey data from a national public opinion poll on the issue of pornography were made available to the Commission by Newsweek magazine. The poll was conducted for Newsweek by the Gallup organization in March, 1985, and involved a sample size of 1020 respondents interviewed by telephone.^[1119]

Comparisons between the Gallup data, where appropriate, will be made with the 1970 Commission survey (see Abelson, et. al., 1970) to examine any observable change.

The 1970 Commission survey used face-to-face interviews from February through April of 1970 with a random sample of 2,486 adults and 769 persons ages fifteen to twenty (Abelson, et. al., 1970). For purposes of comparison with the 1985 sample, only the data from the adult sample for 1970 will be used. The Newsweek-Gallup poll was a telephone survey of 1,020 adults conducted in March, 1985.

The 1970 survey was a far more wide-ranging survey covering a host of areas (including opinions on the effects of sexually-explicit material for which some directly comparable poll data are available from the Newsweek poll), the respondents' experiences with sexually explicit materials, opinions on different categories of sexual explicitness, attitudes toward legal and other forms of control, and attitudes toward different categories of sexual explicitness.

In contrast, the Newsweek-Gallup poll was much more limited, consisting of eight questions. For purposes of additional comparison, a 1977 national Gallup poll provides another trend point which allows comparisons with a 1985 question on the applicability of national versus local standards.

Any comparisons between the 1970 and 1985 findings should be made with caution, given the independence of both surveys and the fact that only a few questions were exactly alike. In those areas where questions were examining similar issues but were not worded the same, only the questions which were more narrowly defined for the 1985 survey were included and any resulting error would be on the side of conservatism. The distinctions between direct and indirect comparisons are carefully noted. A major objective is to note whether patterns observed in 1970 continue in 1985. Comparisons will be made in the three areas: (1) public exposure to sexually explicit materials; (2) perceptions of the effects of pornography; and (3) opinions on the regulation of pornography.

1. Public Exposure to Sexually Explicit Materials

The data from 1970 and 1985 are comparable only in a limited way because of differences in the materials mentioned and changes in technology (e.g., the widespread use of cable and home videos). In 1970, for instance, the respondents were asked if they had "ever seen stag movies or skin flicks." In 1985, respondents were asked whether they had gone to an X-rated movie or bought/rented an X-rated video cassette in the last year. The 1985 respondents were asked if they had "ever read" magazines like Playboy or Penthouse, while 1970 respondents were asked if they had seen or read a magazine "which you regarded as pornographic." Again, we note that this is a loose comparison, only afforded by the fact that the 1985 question is more specific in nature and, therefore, a more conservative estimate.

In response to the question whether they had seen or read a magazine "which you regarded as pornographic," one in five in 1970 said "yes," with twenty-eight percent of the men and fourteen percent of the women responding in the affirmative. However, half of the men and a third of the women in this group were unable to recall the title. Of those titles mentioned, it was clear that the term 'pornographic' embraced a wide variety of material including Cosmopolitan, Esquire, Good Housekeeping and Ladies Home journal (Abelson, et. al., p. 23).

In contrast, two thirds of the 1985 respondents had read Playboy or Penthouse at some time. Over a third said they "sometimes buy or read magazines like Playboy" (37%) while thirteen percent said they "sometimes buy or read magazines like Hustler."

In 1970, fifteen percent of respondents said they had seen a movie they regarded as 'pornographic' in the past year. Again the range of titles mentioned included such films as Butch Cassidy and the Sundance Kid, The Graduate, Easy Rider, and Bonnie and Clyde, in addition to titles that could more likely fall in the "adult" movie category. On the other hand, less than ten percent (7%) of the 1985 respondents had been to an X-rated movie in the past year while close to one in ten (9%) had purchased or rented an X-rated video cassette. The marked difference between the questions asked at both time points precludes any conclusion about any increase or decrease in film viewing in the last fifteen years although the media for purveying adult films certainly has increased.

In 1970 as in 1985, men, younger individuals, and those with more education were more likely to have been exposed to sexually explicit material than women, older respondents, and those less educated (Tables 1 and 2). The differences in exposure between men and women are fairly large both in 1970 and in 1985 but are particularly striking in 1970.

At what age is the average person first exposed to sexually explicit materials? Abelson, et. al. (1970) found that about one in five males and about one in ten females had their first exposure by age twelve. By age seventeen, over half of the males (54%) and a third of the females had been exposed (p. 8). Those exposed earlier also tend to differ from those exposed at a later age. "Young adults, college-educated people, those with relatively liberal attitudes toward sex, and people who have experienced the most erotica recently are all disproportionately more likely than others to have had their first experiences with erotica at a young age" (p. 9).

No comparable age-of-first-exposure question was asked in the 1985 Newsweek-Gallup Poll. A few other studies have similarly examined these questions and the results may identify any changes which have occurred since 1970.

Gebhard (1980) compared data collected by the Kinsey Institute between 1938 and 1960 (using only the data from white males and females with at least some college education-a total of 4,388 respondents) to a much smaller nonprobability sample of undergraduate males and females in one university in 1975. By comparing responses to questions on age and source of first knowledge of such topics as coitus, pregnancy, fertilization, menstruation, and venereal disease. Gebhard concluded that "children and young people are learning the basic facts about sex at considerably younger ages than did their parents and grandparents" (p. 168).

For example, over half of each sex in the 1975 sample knew of coitus by age ten whereas only a third of the earlier sample's females and half of the males had this same knowledge at that age. By age eight, thirty-one percent of the males in the Kinsey sample knew of pregnancy compared to sixty-three percent in the 1975 sample; for females, it was thirty-one percent versus seventy-six percent, respectively.

A second finding of this study was that sources of early sex information appeared to have shifted slightly in relative importance. Same-sex peers remained the major source in both samples but to a lesser degree for the more recent sample, with mothers and the mass media becoming more significant (ranked second and third, respectively). These results, however, are simply suggestive because of the difficulty of generalizing beyond these particular groups of respondents and the limited size of the 1975 sample. These data also gave little indication of whether "mass media" includes pornography.

Another more recent set of data based on a national probability sample of 1071 respondents is available from Canada (Check, 1985). The Canadian results show that adolescents, ages twelve to seventeen, report most frequent exposure of sexually explicit fare. As Table 3 shows, two in five twelve to seventeen year olds view such material in movie theaters at least once a month; over a third (37%) see similar material on home videos with the same frequency.

These results should be viewed with caution because of the small numbers in this age group. The 1970 survey data demonstrated a similar pattern. Respondents in the 1970 sample were asked how many times during the past two years they had seen photographs, snapshots, cartoons or movies of a list of sexually explicit items. Adolescents reported more frequent exposure than adults, with three in ten of the adolescents saying they had seen such material six or more times in the last two years compared to one in four adult males and one in seven adult females.

In comparing his results to the 1985 American Newsweek-Gallup data discussed above for comparable questions, Check found parallel results at least for sexually violent material. Results on nonviolent fare could not be compared because of the differences in question wording. This consistency and the fact that over eighty percent of the sexually explicit material in Canada is from the United States (Special Committee on Pornography and Prostitution, 1985, p. 161) might suggest that the Canadian results may not be dissimilar from what might be found in the United States.

Public Standards of Acceptability

The 1970 Commission survey examined standards of acceptance for various categories of explicitness in two types of media: movies and print. Table 4 shows that there was slightly greater tolerance for sexual explicitness in the print media than in movies (if one compares the percentages of persons advocating total bans on various categories). The print category presents a problem since it does not distinguish between textual and visual or photographic material, which might be found more often in books and magazines, respectively. Restrictiveness also progressively increases the more the behavior departs from what respondents might consider normative. A re-analysis of the 1970 survey data does confirm this observation of acceptability based on perceived normativeness and, in addition, shows that judgments were also related to community size and medium (Glass-man, 1978).

In 1985, slightly different distinctions appear to be made (Table 5). Greater tolerance is shown for film (both theater and video tape cassettes) than for print, with the public more likely to suggest no restrictions for the former. While the survey does not use the wider range of distinctions of sexual activities provided 1970 respondents (a limitation imposed no doubt because of the telephone procedure), the three categories used-nudity, sexual relations, and sexual violence-provide a sufficiently diverse range of themes. The data clearly show greater tolerance for nudity, with a majority maintaining that restrictions should only apply to public display. There was least tolerance for sexual violence, with a majority advocating banning such material. What has been called the "VCR morality" is also very much in evidence here with more than a quarter of the respondents opting for no restrictions on X-rated video tape cassettes. Nearly one in four respondents did not object to the sale or rental of video cassettes featuring sexual violence as long as there is no public display.

These differences are clarified further when one takes into account the respondent's age and gender (Table 6). The young are clearly less opposed than the old, and men more than women, these patterns appearing with fairly high consistency.

There also appears to be some interaction between these demographic characteristics. Greater numbers of older men tend to be more permissive than older women, with about twice as many men over fifty suggesting no restrictions on materials across the board. The gap between men and women narrows significantly among younger respondents (those between eighteen and twenty-nine), with women just as likely as men to favor no restrictions on all materials except magazines with nudity and the sale or rental of videocassettes. Men were more likely to favor no restrictions on these materials than women.

Has there been an increase in permissiveness in the last fifteen years? Again, while some of the categories between 1970 and 1985 are not directly comparable, a reasonable comparison can be made for the category describing depictions of sexual intercourse. For the 1970 sample, only four percent advocated no restrictions on depicting intercourse in books and magazines and the same percentage advocated no restrictions for movies as well. In 1985, twelve percent advocated no restrictions on "magazines that show adults having sexual relations." Twenty percent favored no restrictions on "theater showings of X-rated movies." The assumption we make here, of course, is that most respondents associate X-rated movies with depictions of sexual intercourse but these comparisons are made with this caveat in mind. With the exception of sexual violence in magazines, the percentages opting for no restrictions on various categories of materials are also higher in 1985 than in 1970.

Finally, the 1985 sample was asked whether there should be a single nationwide standard or whether local community standards should be applied. Comparable data collected by the Gallup poll in 1977 provides another data point. As Table 7 shows, respondents in 1985 were almost evenly divided on whether a national or community standard should be used (forty-seven percent versus forty-three percent). The numbers who prefer to see local community standards applied have remained about even in 1977 and 1985-about four in ten respondents. There were as many who indicated standards should be stricter in 1977 as in 1985-fortyfive percent versus forty-three percent. Additional analysis shows that those who indicated standards should be stricter were more likely than those who said standards should be less strict to favor application of a national standard (55% to 41%). Six in ten women were also likely to favor a stricter standard compared to four in ten men.

In the last year, this gap between men and women appears to have increased even more on the issue of restrictiveness. A Washington Post-ABC News survey in February, 1986^[1120] asked the question: "Do you think laws against pornography in this country are too strict, not strict enough, or just about right?" Among men, ten percent said they were about right, forty-one percent said they were not strict enough, and forty-seven percent said they were about right. Among women, on the other hand, only two percent said the laws were too strict, while seventy-two percent-seven in ten women-maintained they were not strict enough. Almost a quarter (23%) said they were just about right.

Perceptions of Pornography's Effects

Respondents are just as likely in 1985 as in 1970 to perceive both positive and negative effects from exposure to or use of sexually explicit materials (Table 8). However, there is a significant increase in the numbers who perceive negative effects from 1970 to 1985. 1970 and 1985 data in this case were directly comparable since the same categories of effects were used.

The most significant changes were in the areas of violence toward women, with the number of respondents indicating these materials could lead some people to lose respect for women increasing from forty-three percent to seventy-six percent and those believing they lead some people to commit rape or sexual violence increasing from forty-nine percent to seventy-three percent in 1985. Whether this reflects greater sensitivity toward women or greater consciousness of sexually violent material available or both is unclear. A slightly greater number in 1985 than in 1970 were also likely to think that sexually explicit materials provide a safe "outlet for people with sexual problems" (34% to 27%) while the number of people who agreed that these could help improve the sex lives of some married couples remained the same (47%).

Public Perception of Pornography as a Social Problem

1970 respondents were asked to name "what you think are the two or three most serious problems facing the country today?" At the height of the Vietnam War, not surprisingly, more than half named this event the most important issue, followed by racial conflict and civil rights, and thirdly, by the economy (36% and 32%). Only two percent said they were concerned about erotic materials. The 1985 Gallup survey asked the question of perceived importance in a different way: by evaluating the problem relative to other social problems and asking the respondent to judge whether progress was being made to solve these problems (see Table 9).

Summary

There is greater overall public tolerance for sexually explicit materials. However, public opinion on restrictiveness clearly differentiates among different media, content depictions, and public access to such materials. There is a preference for no public display of materials featuring nudity and nonviolent sexual activities whereas a majority favor banning materials that depict sexual violence. There is also a greater willingness to impose restrictions on theater showing and magazine publication of sexual activities than on home videos.

The most frequent exposure to pornography is reported by adolescents between twelve to seventeen, a finding reported by the Canadian as well as the 1970 Commission survey. While sexual knowledge appears to be acquired at younger ages, it remains unclear what role pornography plays in this "sex education" process.

Finally, the public perceives both beneficial as well as harmful effects from exposure to sexually explicit materials. Some maintain these materials help improve sex lives of some people, that they provide information about sex, and also provide entertainment. A significant number also feel they lead to a loss of respect for women, a breakdown of morals, and the commission of sexual violence. The changes between 1970 and 1985 are most apparent in the increase in the numbers who perceive that exposure to these materials lead to loss of respect for, and the commission of sexual violence against, women.

Sex Offenders and Pornography

A common contention is that exposure to pornography leads to the commission of sex offenses. There are two ways one can examine this contention: (1) by looking at the relationship between sexual offenses statistics and the availability of pornography, and (2) by examining interview data from sex offenders, investigating the mechanics behind the onset of deviancy and the role of pornography in the commission of sex crimes.

The examination of aggregate social indicators of pornography availability and sexual offense statistics provides another view of the potential relationship between pornography and these offenses. It offers another way of validating results of the laboratory studies or from individual surveys. For example, if the results indicate a higher incidence of sexual aggression in the laboratory studies as a consequence of exposure to particular types of stimuli, and if surveys reveal that individuals who report higher levels of exposure to similar materials also tend to exhibit higher levels of sexual aggression, and if these findings are corroborated with a correlation between aggregate measures of availability and offenses, then we have reason to be more confident in an assertion that exposure to the class of materials in question has a substantial relationship to sexual aggression.

In the case of sex offenders, a comparison of their arousal patterns to those of nonoffender groups is vital, particularly as these patterns correlate with sexual aggression and attitudinal measures. It is reasonable to suggest that findings among nonoffender males who are aroused to coercive sexual themes and who also tend to be more sexually aggressive would be more meaningful if matched by similar patterns among those identified as sex offenders.

From the perspective of the offenders and society as well, understanding their behaviors is crucial because of the social costs in terms of victimization. While the number of sex offenses reported by incarcerated sex offenders appears to be small, results of clinical interviews, conducted with outpatient sex offenders (with great lengths taken to assure confidentiality) reveal that the number of crimes committed by the average sex offender is far greater than generally has been estimated (Abel, Mittelman, and Becker, 1985). Data from two psychiatric clinics obtained from 411 sex offenders revealed a staggering number of multiple victimizations per offender. These offenders attempted an average of 581 sex offenses and completed typically about 533 offenses each, with a mean number of 336 victims each. These attempted or completed offenses were over an average period of twelve years (Abel, Mittelman and Becker, 1985).

Aggregate Indicators: the Incidence of Sex Offenses and Pornography Availability:

One of the most frequently cited studies has been the analysis of sex crimes in Denmark before and after the legalization of pornography in the 1960s (see Kutchinsky, 1973; BenVeniste, 1970). Kutchinsky's data showed a drop in the number of reported sex crimes after legalization and he argued that the availability of pornography is cathartic as it siphons off potentially dangerous sex impulses-the "safety valve theory" (Kutchinsky, 1970, p. 288; Kutchinsky, 1973). Kutchinsky's work was lauded by the British pornography commission (Williams, 1979) for its thoroughness and the restraint with which he interpreted his findings. It singled out the dramatic reduction in offenses against children coinciding with the availability of pornography and, while the Commission did not endorse the "safety valve" hypothesis, agreed that Kutchinsky's interpretation was plausible, absent any other likely factor (p. 84).

On the other hand, Kutchinsky's study and conclusions did not go unchallenged. First, the weight of empirical evidence amassed in the last two decades by social psychologists, particularly in the area of media violence and aggressive behavior, hardly supports catharsis (see Weiss, 1969; Geen and Quanty, 1977; Bandura, 1973; Bramel, 1969; Comstock, In Press; NIMH, 1982).

Second, a number of problems have been raised with Kutchinsky's analysis and interpretations (see Cline, 1974; Bachy, 1976; Court, 1977; Baron, 1984; Malamuth and Billings, 1985). Some of these problems included the lumping together of sex offenses masked a stable, if not an increased, rape rate (Cline, 1974; Court, 1984). Also, such crimes as voyeurism were no longer recorded by police. Kutchinsky (1973) also noted that other activities such as homosexuality were simply tolerated more and certain social changes such as earlier sexual experiences for females meant reduced reports of intercourse with minors (Bachy, 1976).

The problem of using aggregate social indicators such as crime reports is well illustrated not just with reliability problems in reporting, but also in differential use of the data. For example, by Bachy's (1976) review of Copenhagen rape statistics between 1965 and 1974 which showed increases in rape and attempted rape as a proportion of total sex offenses. These offenses included intercourse with minors and indecent exposure, in addition to rape and attempted rape. Court's (1984) analysis of rape statistics for Copenhagen showed a similar upward trend while a fluctuating pattern was demonstrated by Kutchinsky's figures for the same crime in the same city between 1965 and 1970.

More recently, Kutchinsky (1985) has maintained that the increased availability of "hard-core" pornography in Denmark "may have been the direct cause of the real decrease in incidents of peeping and child molestation" (p. 313) and has proposed the "substitution" hypothesis as the most likely explanation. He further cites a similar pattern in West Germany with legalization of pornography in 1973 bringing about a decrease in sex offenses against children. This proposed causal link should be viewed with extreme caution, particularly since pornography availability statistics have not been presented.

Other data are available that allow further cross-cultural comparisons. Abramson and Hayashi (1984), in analyzing pornography in Japan, noted that while it was illegal to show pubic hair and adult genitals in sexually explicit stimuli, pornography appeared to be widely available in this country, including the prevalence of bondage and rape as recurring themes. In terms of rape statistics, however, they concluded that a low incidence of rape appears to be the case and suggested that certain socio-cultural mediating circumstances may be involved. Unfortunately, no data are provided by Abramson and Hayashi on availability or rape rates and at least one study indicates that these rates may actually be increasing. Goldstein and Ibaraki (1983) found that while crime rates have decreased or remained relatively stable among adults, juvenile crime increased from twenty-three percent of all crimes in 1976 to forty-two percent in 1980, occurring mainly in violent crime categories, including rape. The unique character of rape in Japan is also evident from these authors' findings that fifty-seven percent of the total reported rapes are groupinstigated and seventy-five percent are committed by juveniles. Finally, an informal survey reported in this study showed that ninety percent of the women interviewed said they would not report the rape to the police if they had been victimized (pp. 317-318).

Other cross-national data from areas as disparate as England, Australia, Singapore, and South Africa were analyzed by Court (1977, 1982, 1984). His studies compared rape rates in countries or areas where pornography is widely available, and those where restrictions exist. On the basis of his findings, Court advanced the propositions that (1) rape reports have increased where pornography laws have been liberalized, while the same steep rise is not in evidence where restrictions exist; (2) intermittent policy changes or changes in the law are temporarily related to changes in the rape rates; (3) the increase in rape reports does not parallel the increase in serious nonsexual offenses.

While Court's data are intriguing, the case he presents is weakened by (a) the selective use of a small number of countries, and (b) the lack of direct correlational analyses between sexual offense statistics and pornography distribution/circulation figures. The Williams Committee in England (Williams, 1981), in fact, took exception with Court's data, pointing out that he did not take into account the rise of crime in general in England (p. 74) and that the rising trend in rape and sexual assaults started well before what Court determined was the date marking the availability of pornography (p. 76; see Court, 1980, 1985 for responses to the Williams Report). Cocrane (1978) has similarly disputed Court's analysis and interpretations.

Kupperstein and Wilson (1970) of the 1970 Commission staff examined the incidence of sex crimes in the United States and reported that the rise in adult sex crimes (using report and arrest data) was not greater than the rise for other offenses between 1960 and 1969, despite the heightened availability of sexually oriented materials. The two indicators used for the latter were the circulation of Playboy magazine and the number of complaints reported to the United States Post Office for unsolicited sexually oriented mail. The study employed fairly crude measures, simply examining the percentage increase for various sexual and nonsexual offenses.

On the whole, a number of methodological problems characterize some of these early studies: first, the availability of pornography was simply assumed to have increased or decreased following legal changes. Second, direct correlations between the volume of pornography and sexual offense rates were not investigated. Third, sexual offenses were combined, masking important differences between various categories of offenses. Finally, the mediating effects of other variables which could affect the relationship between the circulation of pornography and sexual offense rates were not systematically investigated.

More recently, correlational evidence using more detailed statistical analyses, presents some additional insight into the pornography-sex crimes relationship on the aggregate or societal level in the United States (Baron and Straus, 1985). A fifty-state correlational analysis of rape rates and circulation rates of adult magazines was conducted, using aggregate circulation rates (subscription and newsstand sales per 100,000 population), for eight magazines (Chic, Club, Gallery, Genesis, Hustler, Oui, Playboy, and Penthouse). A fairly strong correlation+.64-was found between these circulation rates and rape rates. This relationship was present even with controls for potential confounding variables such as police practices (measured by police expenditures per capita), propensity to report rape (measured by number of rape crisis centers per 100,000 females; NOW membership per 100,000 females; MS magazine circulation per 100,000 females; and number of battered women's shelters); "southernness" (based on the higher violent crime rates in the South), and "illegitimate opportunities" (referring to greater opportunities to commit crimes in warmer than colder periods; the indicator used was average temperature).

Baron and Straus further found that rape rates are negatively correlated with the status of women when other factors are controlled for. This status-of-women index was measured via economic, political and legal indicators such as women's median income as a percentage of men's; the percentage of female members in the state legislature; and existence of laws giving women the same property rights as men. The study concluded that in a male-dominant society, the lower status of women may be reflected in higher rape rates.

Since it is possible that rape rates also may be a function of the overall culture supporting legitimate violence (that is, the societal endorsement of the use of physical force for socially approved ends, such as crime control or order in schools), the relationship between this factor and rape rates was also examined. Using a twelve-measure index that included such figures as violent television viewing, hunting licenses issued, and use of corporal punishment, no significant association between legitimate violence and rape was found. It is still theoretically possible that rape rates may be influenced indirectly by the level of legitimate violence through the Tatter's inverse relationship with the status of women; that is, cultural support for violence may contribute to sexual inequality which, in turn, may increase the risk of rape.

Finally, the level of social disorganization was also found to be directly related with rape rates and to affect these rates indirectly through its association with the circulation of pornography and the status of women. Other factors found to correlate with rape rates were the extent of urbanization, economic inequality, and unemployment.

In comparing the relative influence of these various explanatory variables, it was found that the proliferation of sexually explicit magazines and the level of urbanization help explain more of the variation in rape rates than social disorganization. The latter is also "more influential" in predicting rape than are economic inequality, unemployment, or sexual inequality. Together, these six explanatory factors explain eighty-three percent of rape rate variations, certainly a considerable proportion of the variance.

A follow-up study by Jaffee and Straus (1986) examined the impact of a variable called "sexual liberalism" on the relationship between these sexually explicit magazines' circulation rates and rape rates. It was hypothesized that a more liberal sexual climate might explain the relationship between sexually explicit magazines' circulation rate and rape by encouraging men to purchase more of these magazines and also encourage more women to report rape to the police. An index based on twenty-two questions in a national survey measuring attitudes toward a variety of sexual issues was utilized as the measure for "sexual liberalism." Results showed that the original relationship between rape rates and circulation rates of sex magazines was non-spurious and that sexual liberalism played a minor role, accounting for only nine percent of the state-to-state rape rate variations. A problem with this study, however, is that it attempts to match individual level measures of attitudes with aggregate-level social indicators, using data from forty states for the former (effectively reducing the original sample size of fifty states by a fifth).

Using the Baron and Straus data set, Scott and Schwalm (1985) essentially confirmed the sex magazine-rape rate relationship although their additional analysis showed that when rape rates were correlated with specific magazines, these correlations were higher for Playboy, Penthouse, and Oui than they were for Hustler magazine. Their contention was that sexual content in Hustler magazine was more likely to be associated with rape since this magazine has more sexually violent material than the other three magazines. Since correlations with the other four magazines were not provided, it is difficult to judge the consistency of such a pattern. Furthermore, such a breakdown is again not very helpful since the level of analysis is aggregate rather than individual. Thus, on an individual level, it will be more meaningful to correlate an individual's scores on sexual aggression measures and that individual's readership of specific magazines; on an aggregate level, it is more appropriate to relate the aggregate offense rate with aggregate availability figures for the material in question. And even on the individual level, there may still be some question as to the actual separability of individual magazine readership. A readership survey conducted for Hustler magazine among its subscribers shows that on average, the typical subscriber reads 3.6 adult men's magazines (Readex, 1984).^[1121]

Scott and Schwalm (1985) also analyzed the effect of three additional variables not investigated by Baron and Straus: the effect of circulation rates on general circulation magazines (e.g., Time, Reader's Digest) and the effect of outdoor men's magazines (e.g., Field and Stream, American Rifleman), the latter using the presumption that an indicator of a "macho" environment could also account for rape rates. Alcohol consumption for each state was also examined. None of these variables was significantly related to rape rates.

Scott (1985) further examined the correlation between adult theaters and rape rates for 1982 and found no relationship to exist. It is quite possible that this finding may be an artifact of the decreasing number of adult theaters in this country as a result of the rise of home videos, as Scott himself pointed out (see also Newsweek, 1985; Knowledge Industries, 1985). He also correlated the number of adult bookstores in each state and rape rates and again, found no relationship. Scott's data may not necessarily be inconsistent with Baron and Straus.' It is quite conceivable that the number of stores may not correlate with rape rates but the actual circulation of the magazines in various outlets do. In any case, Scott's endorsement of the "safety valve" or catharsis hypothesis on the basis of his findings appears premature at the very least.

While Baron and Straus' work is impressive for its methodological care and thoroughness, their findings do not indicate that men are induced to rape as a result of exposure to these magazines. While this is certainly plausible, there are two caveats to their analysis. First, it is a macro-model that is being tested, examining the relationship of various socialcultural factors on rape. Second, given that this is a correlational study, there is always the possibility that there may be some third factor influencing the observed sex-magazine rape rate relationship.^[1122] The crucial causal evidence has to come from an examination of the relationship under controlled conditions, and these studies are discussed below under "Experimental Findings."

On an individual level, some parallel is offered the Baron and Straus data by a recently completed large-scale study on sexual assault among the college student population (Koss, 1986). Correlates of sexual victimization and sexual aggression were examined among 6,000 college students from a probability sample of higher education institutions. This study established a relatively high incidence of sexual assault within this population (336 per 1,000 college women, a rate which includes rape, attempted rape, and forceful sexual contact). The portrait of college men who report behavior that meets legal definitions of rape shows individuals who are sexually experienced, come from homes where family violence was normative, who use alcohol fairly regularly (and reported becoming intoxicated one to three times per month), who regularly discuss with their peers "how a particular woman would be in bed," and who frequently read at least one of the widely available men's magazines.^[1123]

While these results offer correlational evidence, again, they do not support any causal link between readership of such magazines and sexually aggressive behavior. There are a variety of factors that correlate with sexual aggression as this study and the Baron and Strauss (1986) study demonstrate. Both also provide an important contribution towards our understanding of the types of factors, social, cultural, situational, and individual, which interact to explain sexually aggressive behavior as the theoretical thinking behind it.

In the case of causal relationships, the demonstration of a statistical relationship (that is, that the probability of the observed relationship being due to chance is minuscule) is a first requirement. A second requirement is that other competing or alternative explanations have been controlled in order to establish that X indeed causes Y.

In the case of rape rates and circulation rates of adult magazines, establishing a significant correlation between the two is a first step. That such a relationship may in fact be a spurious one due to the existence of some third factor is a second step in establishing the validity of the relationship. Unlike experimental situations, however, where most "alternative factors" are controlled for, by randomly assigning subjects to experimental conditions, one has to be able to identify every potentially significant "third factor" in correlational research and actually account for these in the analysis. Therefore, we find ourselves, at most, in the position of accepting an observed relationship as being plausible but yet cannot fully preclude the possibility of its being spurious.

Evidence from Sexual Offenders

There are three levels by which sex offenders' use of pornography can be evaluated:, first, what are the patterns of their early exposure to pornography? Second, what is the role of pornography in their fantasy and arousal behaviors? Third and most critical, is the question of the role of pornography in the commission of their sex offenses.

Methodological Considerations

A number of elements need to be considered in the study of sex offender populations.

1. Nature of the population evaluated

   The deviant populations most accessible to researchers in the past were incarcerated sex offenders. This category thus constituted the samples described in earlier studies, including the significant pioneering surveys done by Gebhard, et al. (1965) and Goldstein, et al. (1970). There is evidence, however, that data provided by incarcerated offenders tends to vary significantly from non-incarcerated groups (Abel, Becker and Skinner, 1985). The demographic profile of incarcerated offenders, for instance, appears to differ from nonincarcerated groups. For example, Goldstein, et al. (1970) found that while forty-two percent of his control sample has some college education, only twenty-six percent of the rapists, twenty percent and five percent of the male-object and female-object pedophiles, respectively, also had similar educational levels. Gebhard, et al. (1965) similarly found lower educational levels among his sexual offender sample compared to controls. Only thirteen percent of heterosexual child molesters, thirteen percent of homosexual child molesters, and twenty-one percent of rapists had a grade eleven or higher education compared to twenty-one percent of other criminal offenders and fifty-two percent of the control sample. Both these studies examined incarcerated samples.

   Abel (1985), on the other hand, found that among an outpatient sample of 192 child molesters, forty-six percent had at least one year of college, with a quarter of the total sample completing college or having an advanced degree. Marshall's (1985) comparison of eighty-nine outpatient sex offenders with twenty-four control adults showed little difference between the mean IQ's of this group and a comparison control. A mean IQ of 92, 93, 94 and 101 was measured for heterosexual and homosexual child molesters, incest offenders, and rapists, respectively, and 91 for the control sample. It has been estimated that incarceration rates for some sex offenders are low. Only thirteen to sixteen percent of rapists are actually incarcerated, for instance (Abel, Becker and Skinner, 1985; Dietz, 1978), making it likely that an outpatient sample of sex offenders/deviants would more closely resemble the population of deviant cases than an incarcerated one. The representativeness of such an outpatient group still is uncertain, given the fact that these are individuals who, either voluntarily or by court mandate, have sought treatment.

2. Measurement of Arousal.

   An important aspect of evaluating sexual deviance in terms of diagnosis, treatment, and projection of future behavior has been the assessment of arousal patterns. A major weakness in the early studies on sexual deviance was that measures of arousal consisted solely of self-reports. An extensive review of various assessment procedures (Zuckerman, 1971) concluded that the measurement of penis size (penile tumescence) in response to various stimuli provides the most valid indicator of sexual arousal. While the development of the penile transducer provided more accurate assessments of male arousal, problems still exist with this technology. The primary problem is that it is possible for the offender to control his erectile responses (by controlling his attention and sexual fantasies. See Quinsey and Bergersen, 1976; Laws and Holman, 1978; Abel, Becker and Skinner, 1985; Abel, Rouleau and Cunningham-Rathner, in press). However, it has been possible to identify such faked responses under planned treatment situations and to reduce their occurrence but not to eliminate them entirely (Abel, Mittelman and Becker, 1985).

3. Ethical Considerations.

   Clinical researchers are obviously unable to examine sex offenders in laboratory conditions to assess cause-andeffect relationships in the same way their social psychologist counterparts are able to do with non-deviant or "normal" populations. The risks are too great for a group with little or no control over their own behaviors. Furthermore, the notion of informed consent becomes a problem when physiological measures of arousal patterns may reveal interest patterns the patient may not even be aware of (see Able, Rouleau and CunninghamRathner, in press). Other ethical considerations further arise out of the occasional conflicting needs of the judicial system, the offender's needs and rights, therapeutic requirements, and even the public interest (see Bohmer, 1983; Abel, Rouleau and Cunningham-Rathner, in press, for an extended discussion).

A number of important advances have been made in the last fifteen years to elucidate the nature of sexual deviancy, particularly as they relate to the measurement of arousal patterns. On the whole, however, certain inherent limitations exist for this particular population that preclude gaining the fullest knowledge about the antecedents of their sexual behaviors. One of the earliest landmark studies based on interviews with sexual offenders was conducted by the Kinsey Institute for Sex Research (Gebhard, et al., 1965). The study was notable for its scope, including 1365 sex offenders, 888 other criminal offenders, and 477 controls, all white males. The study was conducted during two time periods: 1941 to 1945 and 1953-1955.

Interviews with sex offenders led the authors to conclude that no relation between pornography and sex crimes exists. The researchers, in fact, concluded that the inferior intelligence and education of the average sex offender precludes his deriving sufficient sexual arousal from pornography to lead to overt antisocial activity, a conclusion which has been contradicted by much subsequent data.

Some of the other earlier studies on this question were done for the 1970 Commission. On the basis of these early studies (see, for example, Cook and Fosen, 1970; Goldstein, et al., 1970; Walker, 1970; Davis and Braught, 1970), the Commission concluded that (1) sex offenders did not differ from adults in the general population in their reported immediate responses to reading or viewing erotic materials; (2) that sex offenders generally had less adolescent experience with erotica than the general adult population but did not differ from the latter in adult exposure patterns; and (3) erotica was an insignificant factor in the reported likelihood of engaging in sexual behavior during or after exposure.

Since these early studies, much more has been learned about sex offenders in terms of their arousal patterns and efficacies of various treatment approaches.

Early exposure

Do sex offenders differ from non-offenders in their patterns of early exposure to pornography? Goldstein, Kant, Judd, Rice and Green, (1970) found a high level of exposure to pornography during adolescence among sex offenders (categories in this study included rapists, pedophiles, transsexuals, and homosexuals) but these levels were not significantly different from a nonoffender comparison group. In comparing their samples on exposure to pornography during adolescence, Goldstein and his colleagues found that more rapists than controls had never been exposed to particular types of stimuli. Yet, the differences among the various groups were not statistically significant over the total range of stimuli. Significantly less exposure was reported among rapists to photos of partially and fully nude women and to books describing nudity and oral-genital relations. In fifteen other categories (different themes in different media), the differences were not significant. In their later book elaborating on their findings (Goldstein, Kant and Harman, 1973) Goldstein and his colleagues describe a significantly larger number of rapists as having had exposure to "hard-core" pornography than controls (30% versus 2%) during childhood or between six to ten years old. They also tended to report an earlier age of "peak experience" with pornography, that is, a sexual experience that stood out in their minds the most.

Cook and Fosen (1970) found that among their sample of incarcerated sex offenders and criminal offenders, the latter reported higher rates of exposure to pornography during preadolescence and adolescence. Johnson, et al., (1970), on the other hand, found slightly higher rates of early exposure among a sample of convicted sex offenders who were on probation and receiving therapy compared to the control sample consisting of the respondents in the Abelson, et al. (1970) national survey (44% versus 40%).

Walker (1970) interviewed two groups of male sex offenders, one from a maximum security ward of a state hospital and the second who were prisoners in a correctional facility. Two control groups incarcerated in both facilities for reasons other than sex offenses were utilized in addition to another comparison group of male college students and members of a number of men's service clubs. The latter were more closely matched to the sex offender sample in terms of age.

Data on age of first exposure revealed no overall difference between sex offenders and the combined student and men's club controls. However, portrayals of sexual activities for which the sex offenders had earlier exposure than the men's club control group appeared to be of the more unusual variety: bestiality, group sex, and "sex activities with whips, belts or ropes."

While the student and men's club members had significantly greater exposure to a wider range of sexually explicit depictions than the sex offenders, the latter also had collected pornography for a longer period of time than the men's club members.

Another study conducted for the 1970 commission (Davis and Braught, 1970) found that early exposure to pornography was related to greater involvement in deviant sexual practices among groups of criminal offenders and a comparison group of male students. This was particularly true for what they called "serious deviance," primarily rape. The age-of-exposure variable appears to be crucial as these authors found that exposure to pornography was a strong predictor of sexual deviance among the early age of exposure subjects. They also noted that "exposure to pornography in the 'early age of exposure' subgroup was related to a variety of precocious heterosexual and deviant sexual behaviors."

They found a slightly different pattern among those exposed to pornography at a later age, with the amount of exposure correlated with poor character scores and participation in criminal, deviant, and sexually active peer groups. This result suggests that among those later exposed, such exposure to pornography is part of a deviant and highly active sexual life style. Thus, two separate but related factors-pornography and peer pressure-seem to play some interacting role as sexual behavior patterns develop (Davis and Braught, 1973, p. 194). However, because we do not have age-of-commission data for the more deviant sexual behaviors, a hypothesis that gives a causal status to pornography exposure cannot be supported. Among 476 male reformatory inmates between sixteen to twenty-one years old, a similar association was found between early age of exposure to pornography as well as high exposure and sexual deviance.

Because more recent studies (Abel, Rouleau and Cunningham-Rathner, in press) suggest that over fifty percent of various categories of paraphiliacs had developed their deviant arousal patterns prior to age eighteen, it is clear that the age-of-first-exposure variable and the nature of that exposure needs to be examined more carefully. There is also evidence that the longer the duration of the paraphilia, the more significant the association with use of pornography (Abel, Mittelman and Becker, 1985). On the whole, the conclusion of the 1970 study that "both the extent and frequency of sex offenders' experience with erotic material is substantially less than that of non-sex (criminal) offenders and nonoffender adults during preadolescence and during adolescence" needs to be qualified. These data demonstrate relatively lower levels of exposure among sex offenders when the comparison group is criminal offenders. Compared to "normal" adults, however, the differences appear to be more qualitative than quantitative: sex offenders seem to have been exposed to sexually explicit materials for the first time at earlier ages, and there are some suggestions that the range of material they were exposed to was of the more unusual variety compared to the wider range of materials that control nonoffender groups was exposed to.

It is important to stress that these findings apply specifically to incarcerated samples, particularly groups that were considered serious offenders, given the maximum security facilities housing the Goldstein, et al. sample, the Walker sample and the Davis and Braught sample. A recent study (Carter, Prentky, Knight and Vanderveer, 1985) compared thirty-eight rapists and twenty-six child molesters incarcerated at a state treatment center. No differences were found between the groups in their exposure to pornography in the home (twenty-seven percent of the rapists and twenty-six percent of the child molesters said they had sex materials in their home while they were growing up) and during development (58% of the rapists and 54% of the child molesters had "seen or read sex materials as a teenager"). However, child molesters were found to use pornography more often than rapists in adulthood, were significantly more likely to use these materials prior to and during their offenses, and to employ pornography to relieve an impulse to commit an offense. Because of the absence of a control group of nonoffender adults, it is difficult to determine whether early exposure to pornography in this instance differs significantly from that of a nonoffender sample. The study also does not describe what types of sex materials were involved.

In retrospective interviews with eighty-nine sex offenders (all nonincarcerated and attending an outpatient clinic) and a control sample, Marshall (1985) found that greater numbers in all categories of offenders had been exposed to nonviolent pornography than the comparison group of non-offenders.

The term "pornography" in this case was limited to two categories of materials: "hardcore materials," or "those available only in specialized stores and depicting sexual acts with nothing left to the imagination" (p. 14), and materials depicting "forced sex." These were described to the subjects as those portraying "sexual relations between adult males and adult females where the female displayed a clear unwillingness to participate by both her verbal refusals and her physical attempts to prevent the attack, and the male in the depiction was said to recognize this refusal but ignored it by forcefully enacting his sexual wishes."

Marshall found that over a third of the rapists (35%), two in five homosexual child molesters (41%), a third of the heterosexual child molesters, and only a fifth of the control adults (21%) had been exposed to materials that did not depict forced sex. Only four percent of the rapists and eight percent of the controls were exposed to sexually aggressive portrayals (forced sex) during pubescence. Because of the terse description of "hardcore" sex materials used in this study, it is difficult to reconcile these findings with those of earlier ones suggesting early exposure to depictions of more deviant activities.

It is apparent that these studies cover a variety of comparison situations (no non-offender controls, comparison with nonsexual criminal offenders only), populations (incarcerated, non-incarcerated and in therapy) and a range of measures for early exposure. Certainly, the notion that sex offenders have significantly less exposure to sexually explicit materials than normal adults does not appear to hold for nonincarcerated groups (Marshall, 1985; Johnson, et al., 1970) and, for incarcerated groups, appears to be true when the comparison group is nonsexual criminal offenders. Compared to nonoffenders, rapists differ only on specific types of material (Goldstein, et al., 1970). Only one study (Marshall, 1985) shows somewhat higher levels of exposure than non-offender adults.

Fantasies and Arousal

Studies reviewed by and conducted for the 1970 Commission examined differences in arousal patterns for sex offenders and nonoffenders. These studies showed either that sex offenders were somewhat less responsive than other adults to erotic stimuli (e.g., Gebhard, P. H.,J. H. Gagnon, W. B. Pomeroy and C.V. Christenson, 1965) or that both groups did not differ in their responses to reading or viewing erotic material (Cook and Fosen, 1970; Walker, 1970; Johnson, W.T., L. Kupperstein, and J. Peters, 1970). The Commission concluded in summary that "the available research indicates that sex offenders do not differ significantly from other adults in their reported arousal or reported likelihood of engaging in sexual behavior during or following exposure to erotica." (p. 284)

Later studies have demonstrated that arousal patterns among sex offenders could in fact differ from non-offenders (Abel, Barlow, Blanchard and Guild, 1977; Barbaree, Marshall and Lanthier, 1979, Quinsey, Chapin and Varney, 1981). The 1970 conclusion can be attributed to a number of factors: first, self-report measures of arousal were used for the most part. The problems with reliance on self-reports as the sole arousal measure have already been discussed. Second, many of the studies used stimuli labeled "erotica" without attempting to discriminate among content cues (stimuli used, for instance, were primarily adult heterosexual activities). Finally, with the exception of the Goldstein, et al. (1970) study, differences among sex offenders categories were not examined.

Subsequent studies have shown the importance of discriminating among various categories of sex offenders, content cues, and utilizing physiological measures of sexual arousal.

While other physiological measures have been used as correlates of sexual arousal (e.g., respiration, galvanic skin response, heart rate), these have been viewed as less desirable than direct erection calibration (Zuckerman, 1971) since increases in these variables have also been recorded for other emotional states not related to sexual arousal.

A key study that attempted to distinguish rapists from nonrapists on the basis of erections was conducted by Abel, Barlow, Blanchard and Guild (1977). This study was also important in its attempt to discriminate responses according to consenting and non-consenting stimuli. The development of a "rape index" was another important element in this study. The index was the quotient of the mean percent erection to rape cues to the mean percent erection to mutually consenting intercourse, a measure which was found to have predictive validity in this study and subsequent ones (see Abel, et. al., 1976; Quinsey and Chaplin, 1982; Quinsey, Chaplin and Varney, 1981; Barbaree, Marshall and Lanthier, 1979). The results showed that rapists respond to both rape and mutually enjoyable intercourse cues while nonrapists exhibited arousal only to the latter.

Other studies have similiarly found that rapists show sexual arousal to rape cues as well as to depictions of consenting sexual activity compared to nonrapists who are usually more aroused to the latter (Abel, Becker, Blanchard and Djenderedjian, 1978; Barbaree, Marshall and Lanthier, 1979; Quinsey, Chaplin and Varney, 1981). The nature of sexual cues was further elaborated by Quinsey and Chaplin (1984) who found that rapists did not discriminate among the various sexually explicit narratives used while nonrapists responded most to the consenting sex narratives, less when the sexual partner did not consent, and least when the victim was shown to experience pain.

In comparing these findings to males in the general population, sexual arousal responses have also been found to be indicative of a proclivity to rape but only in combination with other factors will such a tendency be manifested in overtly aggressive behavior (Malamuth, Check and Briere, 1985, Malamuth, In Press).

Child molesters also have demonstrated significantly different arousal patterns with penile circumference measures than a comparison group of non-sex offender patients (Quinsey, Steinman, Bergersen and Holmes, 1975). Twenty male child molesters confined in a maximum security psychiatric institution exhibited significantly higher penile circumference measures when presented with slides featuring children compared to eleven nonsex-offender patients from the same institution and ten control adults from the community.

Marshall (1985) reported that among his sample of eighty-nine sex offenders, two in five of the heterosexual child molesters, two out of three of the homosexual child molesters, and one in two rapists said they used deviant fantasies "usually" or "always" during masturbation. None of the control adults indicated they had these deviant fantasies "usually" or "always" although forty-six percent said they did so "occasionally" or "rarely."

Seven out of eighteen rapists indicated that "consenting pornography" provided a cue to elicit fantasies of forced sex. Similarly, ten of the eighteen who currently used "consenting sex" stimuli used it to elicit rape fantasies.

Abel, (1985) reported that erotica use increased self-reported arousal (i.e., erotica "increased their deviant sexual arousal") more frequently among rapists than among child molesters, with fifty-six percent of the rapists indicating erotica use increased their arousal compared to forty-two percent of the child molesters. Since there were only sixteen rapists compared to 112 child molesters in this report, these findings have to be viewed with caution. In addition, a number of questions can be raised about these data. First, it is unclear what "erotica use" refers to. It could refer to usage for masturbation, for arousal prior to committing an offense, or, perhaps for child molesters, use during the commission of an offense (e.g., to lower the victim's inhibitions). It is also far from clear whether these arousal changes refer to changes in the offender's arousal patterns or whether these are simply their reported reactions to sexually explicit materials. Current evidence suggests a high correlation between deviant fantasies and deviant behaviors (Marshall, 1984; Abel, Rouleau and CunninghamRathner, 1985). Some treatment methods are also predicated on the link between fantasies and behavior by attempting to alter fantasy patterns in order to change the deviant behaviors (Davison, 1968; Marquis, 1970; Marshall, 1973). What is unclear, however, is the use of pornographic stimuli as a precondition for the generation of such fantasies.

Commission of Sex Crimes

Goldstein, et al.'s 1970 data on offenders' and a control group's reaction to a "peak experience" with erotica is reproduced below. "Peak experience" in this instance refers to the most memorable depiction of a stimulus, one "which really stood out in your mind the most" (p. 81). Again, keeping in mind that this sample was an incarcerated sample, the results show that as teenagers, deviants did not differ much from controls in terms of trying to enact the behaviors they had seen. As adults, a quarter of the female-object pedophiles did try the behavior depicted shortly thereafter compared to thirteen percent of the controls, fifteen percent of the rapists, six and seven percent of the homosexuals and transsexuals, respectively.

In Marshall's (1985) sample of eighty-nine sex offenders, slightly more than one-third of the child molesters and rapists reported at least occasionally being incited to commit an offense by exposure to forced or consenting pornography. Pornography as an instigator was not deliberately sought out by every offender in this category to arouse them to offend. For some, pornography as an instigator was simply fortuitous. Fifty-three percent of those child molesters who reported being incited to offend by pornography said their use was deliberate in their preparation for committing an offense, as was the case for thirty-three percent of the rapists. Finally, six of the eight rapists who reported being incited to offend by pornography reported occasional use of "consenting" pornography to elicit rape fantasies which in turn led to the commission of a crime. It is unclear whether the use of this type of material was by choice or because it was the only material available.

Finally, Abel, Mittelman and Becker (1985) evaluated the use of erotica/pornography by 256 paraphiliacs undergoing outpatient assessment-treatment. Regardless of paraphiliac activity, those targeting adults were somewhat more likely to use erotica (60%) than those targeting adolescents (43%) or children (46%).

Categorized according to their primary predispositioning, fifty-six percent of their rapists and forty-two percent of their child molesters implicated pornography in the commission of their offenses.

Again, these comparisons have to be viewed with caution. The disparities in the data can, in part, be accounted for by the questions posed to the respondent and the differences in the samples. In terms of the population differences, Abel's and Marshall's samples are non-incarcerated while Goldstein's sample consisted of incarcerated sex offenders in a maximum security prison. The Goldstein sample was questioned about trying the behavior depicted in the stimulus to which the respondent had recently been exposed, a stimulus "which really stood out in your mind the most" (p. 81). This very specific question regarding the imitation of the most memorable depiction (the "peak experience") likely accounts for the lower figures relative to those obtained in the other studies. The other two studies, on the other hand, used more general questions pertaining to the use of such materials in commission of offenses.

While these figures are suggestive of the implication of pornography in the commission of sex crimes among some rapists and child molesters, the question still remains: is there a difference in the rates of offenses among those who use pornography versus those who don't? The only data available that directly address this issue suggest that these offenses occur regardless of the use of pornography by the offender (Abel, et al., 1985).

Those offenders who did not use pornography did not differ significantly from those who did in frequency of sex crimes committed, number of victims, ability to control deviant urges, and degree of violence used during commission of the sex crime. The longer the duration of paraphiliac arousal, however, the greater the use of pornography.

Based on these data, the authors suggest that sexual deviants appear to come from socially deprived environments which stunt their social and other coping skills. The longer the duration of the paraphilia, or the earlier the onset, the more likely the paraphiliac was to have used erotica. It is difficult to say, however, to what extent this early exposure contributed to the onset of the deviance.

A number of questions are not addressed in the discussion of these data. First, it is not entirely clear what "erotica use" means. Does it mean the offender enjoys viewing the material on a regular basis? Does it mean use for arousal and masturbation? Does it mean use as incitement prior to committing an offense? For a child molester, "use" could refer also to the employment of sexually explicit materials to lower inhibitions of a potential victim and to present behaviors that might be imitated (Russell, 1975). There also appear to be a few inconsistencies in the data. For example, the number of sex crimes of those using erotica (302) is considerably higher than those not using it (224), but the mean number of victims shows a difference in the opposite direction (139 vs. 200). Also, the rationale for the use of a criterion value of p =.001 in combination with multiple t-tests remains unclear.

In testimony before this Commission, Abel (1985) suggested on the basis of these data that sexually explicit materials play an important role in the maintenance of these paraphilias. Greater numbers of deviants report current use of erotica, its use is associated with length of the deviancy, and it appears to play some role in maintaining arousal and masturbatory patterns. As Abel (1985) pointed out, while the use of pornography might decrease the likelihood for some offenders to commit sex crimes in the short run, in the long term, "the pairing or association of deviant fantasies with the pleasurable experience of orgasm perpetuates the deviant sexual interest." It is clear that the role of sexually explicit materials in this maintenance of deviancy needs to be investigated more thoroughly particularly as they relate to repeated offenses.

Summary

While the number of studies on sex offenders has proliferated in the last fifteen years, the etiology of deviancy still remains to be answered.

There is evidence of a correlational relationship between pornography availability and rape offenses in the United States but such evidence remains in need of corroboration by experimental evidence using similar stimuli. Furthermore, correctional data appear inconsistent across cultures. There is little analogous social science evidence on pornography availability and child molestation with the exception of Kutchinsky's recent assertion that increases in availability caused less molestation in Denmark and West Germany (1985). The "causal" assertion here is not only tenuous; clinical evidence of long term use of pornography being correlated with length of the deviancy at least suggests this assertion is debatable.

The contribution of pornography to sexual deviance remains an open question. At present, "no single, comprehensive theory to explain the development of paraphiliac behavior has yet emerged." (Kilmann et al. 1982). Competing models include a psychoanalytic view which views the paraphilia as a symptom of an underlying psychopathology, with its origins in unresolved conflicts during psychosexual development, a Freudian view; a behavioral model which postulates that the occurrence of sexual variance is a result of classical conditioning processes including modeling, reinforcement, generalization, and punishment, much as "normal" sexual behavior also occurs; and a biological model which suggests genetic influences and emphasizes the control of sexual behavior through biological or hormonal means (e.g., Ball, 1968; Berlin, 1983; Money, 1984).

The 1970 Commission's conclusion that sex offenders have less exposure to pornography may have been applicable only to serious sex offenders (that is, those incarcerated in maximum security institutions). At most, a reevaluation of their evidence and those from subsequent studies suggests that rather than frequency of exposure, it may be the quality of that exposure and the age-of-first-exposure that might help explain subsequent sex behavior differences. Malamuth and Billings (1985) have, in fact, suggested that the effect of pornography on rapists may be more pronounced as a function of their more restrictive home environments, with limited or no information on sexuality and male-female relations.

It is unfortunate that the nature of the first masturbatory experiences and the role of pornography in that experience, if any, also remains a gap in our knowledge for future research to address.

Finally, while self-reports of some offenders appear to implicate pornography in the commission of their sex offenses, the objective data of actual offenses committed which show no significant differences between those who use pornography and those who don't have to be viewed as tentative. It is clear that in addition to investigating developmental sexual behavior patterns among offenders, their arousal patterns as these relate to offenses committed should be investigated more thoroughly.

Effects on the "Average Individual": the Experimental Evidence

In order to draw conclusions about whether exposure to pornography leads to or causes certain effects, one would have to look at the experimental evidence for these causal linkages.

The experimental results are presented in terms of effects in the areas of arousal, perceptions, affective states, attitudes, and behavior. Two categories of pornographic stimuli have generally been used to sort out differential effects in these areas: nonaggressive-pornography and aggressive pornography (see, for example, Malamuth and Donnerstein, 1984; Donnerstein, 1983). Some question may be raised about whether in fact these two categories are sufficiently representative of distinctions the average consumer or the public at large might make or whether these two categories afford reasonable conceptual value. Nevertheless, these categories provide a convenient way to organize the results from experimental studies.

The Effects of Violent Sexually Explicit Materials

The findings from studies investigating effects of exposure to sexually violent materials appear to be fairly unequivocal: measures in the areas of attitudes and behaviors have consistently demonstrated changes in attitudes and laboratory-measured behaviors, with the nature of the effect mediated by such additional factors as message cues (e.g., whether the female victim is shown to be abhorring or enjoying the rape) and individual personality differences.

Studies on the effects of exposure to sexually violent material have been conducted primarily in the laboratories of Neal Malamuth (at Manitoba, Canada and University of California, Los Angeles) and Edward Donnerstein at the University of Wisconsin. With the respective colleagues, they have utilized three typical approaches.

The first approach generally has subjects exposed to stimuli (usually varying consent versus force), with physiological penile tumescence and self-report measures of arousal taken during exposure, followed by questionnaires incorporating dependent variable measures (e.g., likelihood of rape, acceptance of rape myths and interpersonal violence, acceptance of sexual violence against women (see, for example, Malamuth and Check, 1980, 1981, 1983).

A second approach typified by Linz (1985) has subjects exposed to one of several types of stimuli over time (neutral, aggressive, or sexually violent of the "slasher" variety) under the guise of a film evaluation study. Prior to this exposure, measures are generally obtained on psychoticism, in part to eliminate participation by subject who might be especially vulnerable to this type of exposure. The second phase has subjects participate in an ostensibly different study in the law school where they are asked to take part in a mock rape trial. Measures are then obtained at this point which assess punitiveness, rape empathy and similar attitudes.

The third approach has been to expose subjects in the laboratory to sexually violent versus comparison material and assess negative effects by utilizing surrogate measures of aggressive behavior (e.g., shock intensities on an aggression machine. See, for example, Donnerstein, 1980; Donnerstein and Berkowitz, 1981).

All three approaches have different virtues which contribute to our ability to understand various dimensions to the problem. For example, the physiological penile measures of arousal provide an independent and objective means of corroborating self-reports. Surrogate measures of aggression avoid the ethical problems of "inducing" actual anti-social behaviors and at the same time can be validated by actual self-reports of aggression in sexual behavior. Finally, the "massive" exposures afford a first step at our efforts to examine the longer-term effects of exposure to sexually-explicit materials.

Effects on Fantasies

Only one study has examined the effects of sexually explicit materials on fantasies. Malamuth (1981) presented two groups of male subjects with a slide-audio show. One version depicted rape and the other showed a mutually-consenting sexual encounter. Analyses of sexual fantasies which subjects were later asked to create and write down indicated that those exposed to the rape version were more likely to create aggressive sexual fantasies.

Aggressive sexual fantasies appear to be fairly common among certain groups of offenders. Gebhard, et al., (1965) found that "patterned rapists" or those who raped repeatedly, were significantly more likely than incidental rapists to often engage in sadomasochistic fantasies (twenty percent versus zero percent). Walker and Meyer (1981) found four in five of their rapists to report primarily deviant sexual fantasies while Abel, Becker and Skinner (1985) similarly reported aggressive sexual fantasies among their outpatient sexual assaulters. What role pornography, particularly violent pornography, plays in the construction of these fantasies remains to be answered.

Effects on Arousal, Perceptions, and Attitudes.

Are there differences in effects from exposure to violent versus nonviolent sexually explicit material? An early study (Malamuth, Reisin and Spinner, 1979) had male and female subjects exposed to one of the above stimuli or a neutral one. The materials presented were pictures from Playboy or Penthouse magazines for the sexual exposures and from National Geographic for the neutral exposure. Sexually violent depictions included pictures of rape or sadomasochism whereas the sexually nonviolent material had no aggressive elements. After viewing the materials, subjects filled out a mood checklist. This was followed ten minutes later by an assessment of reactions to rape after the subjects had viewed a videotaped interview with an actual rape victim as well as an assessment several days later in an ostensibly different study. Both types of stimuli were found to reduce the extent to which subjects perceived that pornography may have detrimental effects but neither one affected reactions to rape. Correlational data, on the other hand, showed that sexual arousal to the sexually violent depictions were significantly related with a selfreported possibility of engaging in rape.

Another study (Malamuth, Haber and Feshbach, 1980) examined the effects of written descriptions of a sexual interaction based on a feature from Penthouse magazine and modified to created a violent and nonviolent version for male and female subjects. In this study, males who had been exposed to the sexually violent depiction (sadomasochism) perceived more favorably a rape depiction that was presented to subjects subsequently. Subjects were found to believe that a high percentage of men would rape if they knew they would not be punished and that many women would enjoy being victimized. Finally, of the fifty-three male subjects, seventeen percent said they personally would be likely to act as the rapist did under similar circumstances. Fifty-three percent of these males responded similarly when asked the same question if they could be assured they would not be caught.

In order to draw out the various dimensions in the portrayals of sexual violence which might explain the exhibition or inhibition of sexual responsiveness, Malamuth, Heim and Feshbach (1980) conducted two experiments on male and female students. The first experiment replicated earlier findings that normal subjects seem to be less aroused by sexual violence than by "nonviolent erotica." A second experiment manipulated reactions of the rape victim with one version showing her as experiencing an involuntary orgasm and no pain. The second version had her experiencing an orgasm with pain. Both male and female subjects were aroused to these depictions, with female subjects more aroused by the orgasm with no pain version while the males were most aroused by the orgasm with pain stimulus. The authors postulated in this case that under certain conditions, rape depictions can be arousing, particularly when the rape victim is shown experiencing an orgasm during the assault. According to the authors, subjects may have reinterpreted the events preceding the depiction of the victim's arousal so that the rape is now viewed as one that is less coercive and less guilt-inducing.

Three additional studies (Malamuth and Check, 1980a, 1980b, 1983) provide further evidence that victim reactions have a significant impact on sexual arousal and behavioral intentions. Results from one of these studies showed that both male and female subjects exhibited higher arousal levels when portrayals showed an aroused female, regardless of whether the context was a rape or a mutually consenting situation. The second study (Malamuth and Check, 1980a) similarly showed that male subjects had higher penile tumescence scores when viewing a victimaroused rape portrayal compared to a portrayal showing victim abhorrence. Significant correlations were also obtained between the reported possibility of engaging in similar behavior, sexual arousal to rape depictions and callous attitudes toward rape.

The effect of sexually violent depictions on attitudes has also been demonstrated with male and female subjects reporting greater acceptance of rape myths after exposure to such material (Malamuth and Check, 1980a; 1985; Malamuth, Haber and Feshbach, 1980).

In an attempt to approximate a "real world" situation, Malamuth and Check (1981) had male and female subjects view full-length features as part of campus cinema showings. The films-Swept Away and The Getaway-represented sexually violent films whereas control subjects viewed a nonviolent feature film. Dependent measures were obtained after a week in a questionnaire presented as a separate sexual attitudes survey. These measures included rape myth acceptance measures, measures on the acceptance of interpersonal violence as well as adversarial sexual beliefs, measures developed by Burt (1980). Results showed that exposure to sexual violence increased male subjects' acceptance of interpersonal violence against women. A similar trend, though statistically nonsignificant, was found for the acceptance of rape myths. There were nonsignificant tendencies for females in the opposite direction. In addition to the advantage of external validity from this field experiment, the problem of demand characteristics in some laboratory experimental situations is quite effectively dealt with in this study.

Aggressive Behavior

Donnerstein (1980) had male subjects provoked or treated in a neutral manner by a male or female confederate, then had them view one of three films: a sexually explicit film, a film depicting a rape, and a neutral film. Results of this study show that when the target of angered subjects was a male, there was no difference in aggressive behavior (measured by shock intensity on an aggression machine) among males in the erotic and the aggressive-pornographic conditions. However, when the target was a female, aggressive behavior was higher only in the aggressive pornographic film condition, regardless of provocation.

To account for the impact of victim reactions in a rape portrayal, Donnerstein and Berkowitz (1981) had male subjects angered by a male or female confederate. Following instigation, they then watched one of four films: a neutral film, a non-aggressive pornographic film, an aggressive pornographic film with a positive outcome (where the woman is smiling and offering no resistance, becoming a willing participant in the end) and the last with a negative outcome, where the woman is shown exhibiting disgust and humiliation. Subjects who were angered by a male confederate were not significantly more aggressive towards the male instigator after viewing the pornographic or aggressive-pornographic film; those angered by a female, however, showed significantly higher levels of aggressive behavior in both aggressive-pornographic conditions, that is, those that portrayed a negative and those showing a positive outcome.

What about the effects of positive and negative outcomes on non-angered subjects? The same study (Donnerstein and Berkowitz, 1981) examined this issue using only female confederates. Results showed that for non-angered subjects, only the aggressive-pornographic film with a positive ending elicited higher aggression levels. Subjects exposed to this version also saw the woman portrayed as suffering less, enjoying more, and being more responsible for her situation. These findings suggest the importance of dis-inhibiting factors that might produce a readiness to respond (e.g., anger or frustration) and message cues (e.g., enjoyment of sexual coercion) as enhancing the likelihood of laboratory aggressive behavior. These are also short-term effects although with appropriate cues, there might be long-term effects as well. This remains speculative at this point (Malamuth and Ceniti, 1986).

A recent study demonstrates that such laboratory aggression is not always manifested when these "enhancing" factors are absent (Malamuth and Ceniti, 1986). Two groups of subjects were exposed to either sexually violent or sexually nonviolent depictions in movies, books and magazines over several weeks and compared to a third no-exposure control group. Several days later, in what was presented as a different study on ESP, measures of laboratory aggression using aversion noise were obtained in the typical aggression paradigm. No differences were found among the three exposure conditions. The authors speculated that a more immediate measure, incombination with stimuli which "prime" thoughts and feelings relevant to the exhibition of specific behaviors might be more conducive to an individual's performance of such behaviors.

An important study that clarifies the interaction of motivational, message and inhibitory factors as predictors of self-reported sexual aggression (Malamuth, In Press) has demonstrated that (a) such factors as hostility to women, dominance and acceptance of interpersonal violence, arousal to sexual violence, and sexual experience all correlate with sexually aggressive behaviors; (b) the occurrence of these aggressive behaviors is better "explained" or "predicted" by these factors in combination; (c) arousal to sexual aggression correlates with dominance and hostility to women and is also an important predictor of sexual aggression; and (d) these self-reports of sexually aggressive behavior are also correlated with laboratory measures of aggression.

Effects of Massive Exposure

In a study designed to evaluate the effects of massive exposure to sexual violence and to further explore the components of the desensitization process, a series of four studies-all part of a Ph.D. dissertation were conducted. (Linz, 1985). College males were exposed to a series of "slasher films," all R-rated, using a formula of sexual explicitness juxtaposed with much blood and gore. A typical example is a scene from Toolbox Murders showing a naked woman taking a tub bath, masturbating, then being stalked and killed with a power drill by a masked male. Comparisons were also made among R-rated nonviolent films and X-rated nonviolent films, both of which included sexually explicit scenes (the former were of the teenage sex films variety).^[1124]

After viewing one film per day for five days, subjects were asked to participate in what was presented as a different study-a pretest of a law school documentary-then completed a questionnaire assessing the defendant's intentions, the victim's resistance, responsibility, sympathy, attractiveness, injury and worthlessness.

Among his findings:

• Those who were massively exposed to depictions of violence against women came to have fewer negative emotional reactions to the films, to perceive them as significantly less violent, and to consider them significantly less degrading to women.
• This desensitization appeared to spill over into a different context when asked to judge a female victim of a rape. Those massively exposed to sexual violence judged the victim of the assault to be significantly less injured and evaluated her as less worthy than did the control group.
• There were no differences between subjects exposed to the teenage sex film or the X-rated film and the control group on either pretrial measures on objectification of women, rape myth acceptance or the acceptance of conservative sex roles or on the post-trial measures (defendant guilt, verdict, victim responsibility).
• Two movies (about three hours viewing time, about twenty to twenty-five violent acts) were sufficient to obtain a desensitization effect similar to the effect obtained after exposure to five movies, suggesting that desensitization can occur fairly rapidly.
• These findings were most pronounced for those subjects high on psychoticism and exposed to the highly sexually violent film. These individuals were significantly more likely to endorse the use of force in sexual relations and to evaluate the victim portrayed in the rape case as less credible, less worthy, and less attractive.

The effectiveness of debriefing procedures was assessed and the measures were found to be generally effective in reducing negative effects observed after film exposure.

Krafka (1985) used these same R-rated "slasher" films in a study similar to Linz's but using female subjects. Krafka also used these films as stimuli for a "violent" condition and contrasted this with exposure to sexual violence and to an X-rated set of films. The effects of massive exposure obtained for male subjects were absent for females.

It is clear that for males, exposure to sexually explicit materials juxtaposed with violence directed at a female target enhances callous attitudes in similar situations involving women as victims.

The Effects of Nonviolent Sexually Explicit Materials

The importance of specifying various contingent conditions under which certain effects may or may not be obtained becomes immediately obvious when one looks at the findings in this area. It is also clear that while there are a greater number of studies that examined the effects of nonaggressive sexually explicit materials, particularly if one includes the 1970 Commission studies, the diversity of dependent variable measures as well as experimental stimuli used is also greater than those in the area of sexual violence.

A number of different effects from a variety of studies have been obtained in the areas of affect, attitudes as well as behavior.

Affective and Perceptual Responses

Wishnoff (1978) exposed sexually inexperienced undergraduate females to explicit erotic films. He found that sexual anxiety decreased while expectations about engaging in intercourse in the near future increased significantly.

Along the same lines, Byrne (1977) and Byrne and Byrne (1977) suggested that initially, exposure to sexually explicit materials may offend and disturb some, or produce apprehension in others. These authors then hypothesized that frequent exposure reduces negative reactions and negative appraisals of these reactions. Once tolerance increases, the stimuli leads to greater pleasurable sexual fantasies and greater enjoyment, a hypothesis generally supported by their data.

Perceptual judgments have also been demonstrably affected by exposure, particularly in the areas of comparative judgments and estimations of reality. Kenrick and Gutierres (1980) found subjects' judgments of the attractiveness of an average female were lowered by exposure to media females. Proposing that such effects could be more significant in the realm of sexually explicit materials, Gutierres, et. al., (1985) did a follow-up recently in which subjects were asked to assess characteristics of others after exposure to slides of Playboy and Penthouse models. In four successive experiments, target persons rated were a stranger and the subject's spouse or long-term live-in partner. Both types of target persons were more negatively rated only by male subjects. Similar results were obtained after males were exposed to "beautiful females in sexually enticing activities" (sexually provocative poses or precoital and coital activities) in contrast to males exposed to less attractive females (Weaver, Masland, and Zillman, 1984).

This perceptual contrasting of aesthetic appraisals is contingent on whether the rated target and the comparison target are associated (Melamed and Moss, 1975; Griffitt, 1971). For example, when an individual is presented in the context of attractive friends, that individual tends to be rated as more "attractive." In the case of comparing media models with a significant other, on the other hand (where presumably there is no association between the target and the comparison), the comparison stimulus, or the media model in this case, "provides an anchor or contrast point for the evaluation of the target stimulus." (Melamed and Moss, 1975, p. 129).

Hatfield and Sprecher (1983) exposed males to "a Playboy-type article-a romantic seduction scene designed to be arousing." They predicted that a sexually aroused male would exaggerate a woman's sexual desirability as well as her sexual receptivity. Male subjects were then shown a photograph of "a potential date." Both predictions were confirmed. Aroused men, according to the authors, were more likely to agree that their potential date was "amorous," "immoral," "promiscuous," "willing," "unwholesome," and "uninhibited."

Different results were obtained by Dermer and Pyszcynski (1978) in an investigation of the effects of erotica on males' responses to women they loved. They were particularly interested in whether erotica would enhance "loving" or "liking" responses. Males who read an erotic story (an explicit account of sexual behaviors and fantasies of a college female) reported greater romantic involvement than those in a control condition. That is, they were more apt to report expressing "loving" than "liking" statements to their loved ones when sexually aroused than when not sexually aroused.

In looking at the above studies as a whole, it is quite possible that with "loved ones," could accentuate perceptual judgments while stimuli that primarily enhance arousal reactions (as in the Hatfield and Sprecher, 1983 and Dermer and Pysczynski, 1978 studies which used textual material) enhance more "loveoriented" responses for loved ones and "lustoriented" responses in a dating situation.

Effects on Behavior

Initial studies conducted for the 1970 Commission showed that sexually explicit materials had either no effect on sexual behavior or when effects were observed, there were generally slight increases in those sexual activities already in the individual's established repertoire (Amoroso, et. al., 1970; Byrne and Lamberth, 1970; Kutchinsky, 1970). These behavioral effects generally occurred within a short period after exposure. However, as one of the 1970 research investigators observed, it was also possible that,

the effects of erotica on behavior could have been obscured in the initial body of research because two major components of the influence process were missing from the early investigations: the extended time period necessary for change to occur and the specification of the depicted behavior as well as the relationship between the interactants (Byrne and Kelley, 1984).

While more recent studies examined the impact of nonviolent sexually explicit materials after repeated exposure, others have also examined behavioral effects after short-term exposure. It is in the latter area of behavioral effects from exposure to nonviolent sexually explicit stimuli where apparently conflicting results are found.

Baron and Bell (1977) exposed male students to stimuli that included semi-nude females, nudes, heterosexual intercourse and some explicit erotic passages. The mild erotic stimuli (semi-nudes and nudes) inhibited aggression levels whereas the "stronger" stimuli had no effects. A follow-up study (Baron, 1979), this time on female subjects, using the same stimulus materials found mild stimuli inhibiting aggressive behavior while the stronger stimuli increased aggression. Both these studies measured aggressive behavior via "shocks" delivered on an aggression machine.

In another study, photographs variously depicting "nonerotica," nude females, and couples in sexual activities were shown to male subjects (Zillmann and Sapolsky, 1977). Additionally, subjects were either provoked or unprovoked. For the latter group, no differences in aggression levels by type of stimulus were observed. No differences were observed in aggression levels for subjects who were provoked either, although respondents in this condition also exhibited lower annoyance levels. The authors explained these findings in terms of the aggression-reducing effect of relatively non-arousing but usually pleasant sexually explicit images which act to reduce annoyance or anger and consequently, aggressive behavior.

Along these lines, Sapolsky (1984) has suggested that content characteristics have an impact on affective states (that is, how pleasing or displeasing the stimulus is) as well as on arousal levels. The combination of these factors appear to produce differential responses.

Situational factors such as provocation and the removal of restraints against aggression appear to further mediate the effects of nonviolent pornography on viewers. Donnerstein, Donnerstein and Evans (1975) found that "mild erotica" (semi-nudes and nudes from Playboy) inhibited aggressive responses in contrast to "stronger erotica" (frontal heterosexual nudes in simulated intercourse and oral-genital contact) which enhanced aggression, particularly for previously provoked subjects. A subsequent study similarly showed that a pornographic film (black and white stag film depicting oral, anal intercourse and female homosexual intercourse) increased aggression levels among angered males to a significantly greater extent than a neutral film (Donnerstein and Barrett, 1978).

In comparing the effects of both aggressive and erotic films on aggressive behavior of male subjects, Donnerstein and Hallam (1978) found both types of stimuli to increase aggressive behavior against both a male and a female target. However, when these subjects were given a second opportunity to aggress, these responses increased in the pornographic film condition for the female but not for the male target. The second aggression opportunity, the authors suggest, acts to reduce restraints on aggression against women.

In sum, the experimental effects from exposure to nonviolent pornographic material appear to be mediated by a number of conditions: the strength of the stimulus to induce arousal, the affective nature of the stimulus, and situational factors such as the removal of restraints against aggression.

Effects from Longer Term Exposure

A number of studies, both from the 1970 Commission and more recent ones, examined the effects of "massive" exposure to pornography. "Massive exposure" in these studies means exposure over a duration of one to several weeks. Mann, Sidman and Starr (1970) exposed married couples in four consecutive weekly sessions to sexually explicit films or to nonerotic films (for the control group). Sexual activities were recorded in diaries by the subjects during the exposure period and attitudes toward pornography also assessed both prior to and after exposure. Sexual activities increased in frequency during exposure days although these activities were ones these subjects normally engaged in (i.e., they were not related to specific ones portrayed in the stimulus materials). An additional finding was that the reported stimulating effect grew weaker as the weeks progressed. Whether this diminution is attributable to boredom or to habituation is not entirely clear.

Howard, Reifler and Liptzin (1971) similarly exposed male college students to heavy doses of pornographic films, photographs, and reading material during ninety-minute sessions over a three-week period. Experimental subjects could choose from among these materials and other "nonerotic" ones during the first ten sessions. This was followed by three sessions where the original pornographic material was replaced by new ones. During the last two sessions, the "nonerotic" materials were taken away. Control subjects were not exposed to these types of materials. The findings, based on physiologic and attitudinal measures, revealed initial high interest which faded rapidly with repeated exposure. After this period of unrestricted exposure, the provision of new materials failed to revive interest. Decreased penile response was measured as well as concomitant reductions in other responsiveness measures (e.g., heart rate, respiration rate and skin temperature). While the authors interpreted these results in terms of boredom, Zillman and Bryant (1984) suggested that habituation is a potential alternative explanation based on the premise that continued exposure to emotion-inducing stimuli produces declines in the arousal component of the reaction: evidence that habituation effects might be occurring.

To test this hypothesis, Zillman and Bryant (1984) had eighty male and female undergraduates randomly assigned to a massive, intermediate, no exposure or control group. Subjects in the three experimental groups met in six consecutive weekly sessions and watched six films of eight minutes duration each, with varying degrees of exposure to the explicit sex films. Ostensibly, the subjects were to evaluate the aesthetic aspects of these films. All erotic films depicted heterosexual activities, mainly fellatio, cunnilingus, coition, and anal intercourse, none of which depicted infliction of pain. The nonerotic films were educational or entertaining materials, all previously judged as interesting. Experimental subjects returned to the laboratory one week after treatment and were then exposed to three films of varying degrees of explicitness (pre-coitus, oral-genital sex and intercourse, and sadomasochism and bestiality) followed by measurements of excitation levels (heart rate and blood pressure) and affective ratings.

Two weeks after initial treatment, subjects were randomly assigned (within initial exposure treatments) to view one of the following: (a) a film depicting oral-genital sex and heterosexual intercourse; (b) a film depicting sadomasochistic activities; (c) a film featuring bestiality; (d) no film. Measures of aggressive behavior also were obtained at this point.

The results three weeks later indicated that with increasing exposure to various explicit stimuli, arousal responses diminished, as did aggressive behavior. Furthermore, more unusual or "harder" erotic fare appeared to grow increasingly more acceptable with subject evaluations that the material was offensive, pornographic or should be restricted progressively diminishing. Measures of sex callousness suggested further habituation effects as did projective measures of the commonality of these behaviors. According to Zillman and Bryant, these effects were, "evident for both male and female subjects." Similar habituation effects after "massive exposure" were reported by Ceniti and Malamuth (1984) for subjects who were "force-oriented," effects which were most pronounced with exposure to sexually violent depictions. Arousal patterns were not affected, however.

An earlier report on other aspects of the same study (Zillman and Bryant, 1982) showed that subjects also exhibited greater sex-callousness, using measures developed by Mosher (1970). They also showed some cognitive distortion in terms of exaggerated estimates of the prevalence of various sexual activities as a result of massive exposure.

There is contrary evidence from Linz (1985) on the effects of massive exposure to nonviolent sexually explicit materials in a study described earlier under Effects of Massive Exposure to Sexual Violence. Subjects exposed to R-rated "slasher" films, "teen sex" films and "X-rated nonviolent films"^[1125] did not show the same effects in a rapejudgment situation as did the "slasher" films which showed perceptual changes described as desensitization to film violence and to violence against women.

Another investigation into the effects of massive exposure to nonviolent sexually explicit materials tested the habituation hypothesis (Zillman and Bryant, in press) using both male and female students and adults from a metropolitan community similarly examined effects of massive exposure. This time, the "behavior" of interest was choice of entertainment material. Two weeks after exposure, subjects were provided an opportunity to watch videotapes in a private situation with G-rated, R-rated and X-rated programs available. This opportunity to view was provided during an ostensible "waiting period" between procedures, with the subject's choice of entertainment and length of viewing unobtrusively recorded. Subjects with considerable prior exposure to common, nonviolent pornography showed very little interest in this type of fare, choosing instead to watch more uncommon materials that included bondage, sadomasochism, and bestiality. These effects, while observable among both males and females, were again more pronounced among the former.

While habituation is certainly a plausible explanation for these findings, choice of entertainment fare on the basis of stimulus novelty cannot be precluded entirely (see Kelley, In Press). An examination of the mean amount of time spent viewing the video tapes shows that for those massively exposed, male students watched an average of three and a half minutes of "uncommon fare" (featuring bondage, S & M, bestiality) while female students watched an average of a minute and a half, with viewing times for their nonstudent

counterparts only slightly higher. Keeping in mind that subjects had fifteen minutes of viewing time, the graduation to a preference for stronger fare, or habituation, does not seem to be firmly supported by the data. Furthermore, the measurement situation might also be viewed as "permission-granting," with choice of what might normally be considered taboo material being more permissible or socially condoned. One could argue that greater availability of these materials in the real world might also be analogous to an indication of social sanctions being lifted, so to speak, and the laboratory evidence obtained here certainly merits more attention, perhaps through longitudinal studies.

Further measures were obtained from the same samples of subjects in the last study described above in the areas of "sexual satisfaction" and "family values," both through an extensive battery of questions (Zillman and Bryant, 1986a, 1986b). Subjects were asked how satisfied they were with their present sexual partner, their partner's physical appearance, affectionate behavior, commitment and so forth. Their findings showed significantly increased dissatisfaction in these various areas of sexuality after massive exposure.

In the area of "family values," a variety of questions tapped attitudes on pre-marital and extra-marital sex, estimations of occurrences of "sexual faithfulness" in the population, and perceptions of the institution of marriage and divorce. Again, massive exposure appears to have increased acceptance of premarital and extra-marital sex and diminished the importance of the institution of marriage. These findings have to be viewed with caution since the large number of statistical tests conducted increases the chances of obtaining false positive conclusions. Because of the complexity of the experimental procedures, the long battery of questions asked, and the absence of a measure validating the effectiveness of the cover story, we must also view these findings as tentative and worthy of further examination.

On the basis of the above findings, it appears that short-term effects have been observed in the laboratory but under very specific conditions. These conditions should be further elaborated on in future research. Massive exposure studies varying the lengths of exposure, on the other hand, suggest that certain types of effects may occur with long-term exposure. The question arises whether this is true of all types of sexually explicit stimuli that do not have any violent elements.

A recent Canadian study has tried to address this issue (Check, 1985). Four hundred thirty-six college students and nonstudent metropolitan Toronto residents recruited by means of advertisements, were exposed over three videotape viewing sessions to one of three types of materials, or to no material at all. The stimulus materials were constructed (primarily because no materials could be found that exclusively contained the intended manipulations) from existing commercially available entertainment videos to represent one of the following:

1. Sexual violence-Scenes of sexual intercourse which included a woman strapped to a table and being penetrated by a large plastic penis.
2. Sexually explicit and degrading-Scenes of sexual activity which included a man masturbating into a woman's face while sitting on top of her.
3. Sexually explicit-Sex activities leading up to intercourse between a man and woman.

These categorizations were validated in preliminary questionnaires assessing subjects' perceptions of these materials. Results indicated that exposure to both the sexually violent and the nonviolent dehumanizing pornography (1) were more likely to be rated "obscene," "degrading," "offensive" and "aggressive;" (2) tended to elicit more pronounced feelings of anxiety, hostility and depression; and (3) tended to be successfully differentiated from the materials classified as "erotica." The patterns were less clear on reported likelihood of rape measures and reported the likelihood of engaging in coercive sex acts. While those in the violent and in the degrading exposure conditions reported significantly greater likelihood of engaging in these behaviors compared to the control group, an effect more pronounced among those with high pyschoticism scores, those exposed to the "erotica" stimulus did not differ significantly from either the control or both pornography conditions. The findings also have to be viewed with caution as the exposure conditions were not completely equivalent (i.e., the no-exposure control group came in for a single session while the experimental groups came in for four sessions), a caveat Check recognized and discusses. Finally, it is not entirely clear what differential effects on the exposure groups the preliminary instructions to all subjects might have had which included some reference to the study being funded by the Fraser Commission on Pornography.

Similar findings were obtained by Senn (1985) for female subjects exposed over four sessions to slides of "erotica," "nonviolent dehumanizing pornography," and "violent pornography." The first class of materials were discussed as mutually pleasurable sexual expression between two individuals presented as equal in power. The second category was described as having no explicit violence but portraying acts of submission (female kneeling, male standing; female naked, male clothed) while the third included acts of explicit violence in the sexual interaction (e.g., hair-pulling, whipping, rape).

Both violent and nonviolent pornography resulted in greater anxiety, depression and anger than erotica and both were also reliably differentiated from the latter on a number of affective dimensions, with "erotica" consistently rated more positively.

These findings on non-violent, "degrading" pornography are by no means definitive but they do suggest the importance of examining the effects of various content attributes.

Individual Differences

Not everyone reacts in the same way to sexually explicit materials. Researchers have examined various individual explanatory variables which might explain more fully why individuals respond in different ways. We do not intend an exhaustive summary of the variety of individual attributes examined but merely wish to illustrate that observed effects are mediated by a number of factors. Three sets of factors will suffice for discussion.

One characteristic which has been examined is gender. It has often been asserted that females are less interested in sex than males. Some of the early studies in sexual behavior (Kinsey, Pomeroy, Martin and Gebhard, 1953) concluded that females were disinterested in pornography and were less aroused by it. The same sex differences were reported in the national survey of the 1970 Commission (Abelson, et al., 1970).

Experimental findings, however, seem to suggest otherwise. Males and females in laboratory-exposure situations reported the same levels of arousal in response to sexually explicit stimuli (Sigusch, et al., 1970; Byrne and Lamberth, 1971; Griffit, 1973). Females, however, are also more apt to report negative affect toward erotic stimuli, that is, they report more shock, disgust, and annoyance than males (Schmidt, et al., 1973). These differences, not surprisingly, are even more pronounced when aggressive sexual themes such as rape portrayals are employed (Schmidt 1974). The context of the portrayal is also significant as Stock (1983) demonstrated. Female subjects exposed to an eroticized version of a rape exhibited high arousal levels while a version which emphasized the victim's fear and pain elicited negative affective reactions and lower arousal levels. Krafka's (1985) female subjects did not exhibit the same negative effects that Linz's (1985) males did after exposure to R-rated slasher films which the former attributed to some emotional distancing because the victim in these films was invariably female.

Personality differences also mediate effects. One personality dimension which has been examined is "psychoticism" (Eysenck and Eysenck, 1976) which Barnes, Mamaluth and Check, 1984a, 1984b) found to be positively related to the enjoyment of force and unconventional sexual activities. Linz (1985) and Check (1985) similarly found psychoticism scores to be highly correlated with the acceptance of rape myths.

Finally, experiential factors also help explain response differences. Those with more previous experience with sexually explicit materials also tend to be less inclined toward restrictions (Newsweek-Gallup Survey, 1985) and also tend to exhibit more sex-calloused attitudes (Malamuth and Check, 1985) and more self-reported sexually aggressive behavior (Check, 1985).

Summary for Violent and Nonviolent Sexually Explicit Materials

In evaluating the results for sexually violent material, it appears that exposure to such materials (1) leads to a greater acceptance of rape myths and violence against women; (2) have more pronounced effects when the victim is shown enjoying the use of force or violence; (3) is arousing for rapists and for some males in the general population; and (4) has resulted in sexual aggression against women in the laboratory.

Malamuth's (In Press) research has further demonstrated that such attitudes as rape myth acceptance and acceptance of violence against women are correlated with arousal to such materials and with "real-world" sexual aggression and that subjects who have demonstrated sexual aggression in the laboratory are also more likely to report using coercion and force in their actual sexual interactions. The validation of the measures used in his studies, the use of physiological measures of arousal, and the attempt to systematically examine patterns among different populations with a variety of measures, arousal, attitudinal and behavorial, all tend to provide the type of convergent validation we feel is required of social science evidence.

We are less confident about the findings for nonviolent sexually explicit materials and we hasten to add that this is not necessarily because this class of materials has no effects but because the wide variety of effects obtained needs to be more systematically examined and explained. We can speculate, as have others, about potential explanations regarding some of these differences. For example, Check and Malamuth (1985) have pointed to the differences between Mosher's (1970) lack of effects on sex callousness and Zillmann and Bryant's (1982) finding of greater sex callousness from exposure to nonviolent sexually explicit stimuli (using the same scale developed by Mosher) as possibly attributable to a difference in stimulus characteristics. Mosher's film, based on his own descriptions, depicted "more affection than is typical of much pornography,", while Zillmann and Bryant's (1984) material tended to portray women as "nondiscriminating, as hysterically euphoric in response to just about any sexual or pseudosexual stimulation, and as eager to accommodate seemingly any and every sexual request" (p. 22). Check and Malamuth (1985) maintain that the portrayal in Zillmann and Bryant's study suggests "a dehumanized portrayal of women, which had the effect of generating disrespectful, anti-female attitudes in both male and female subjects" (p. 205).

This explanation could conceivably hold for the differences between Linz's (1985) findings and those of Zillmann and Bryant (1984). Because specific attributes that may characterize these films (other than the fact that they contain no violence) and explain their effects are either confounded (i.e., more than one factor is emphasized, making it difficult to attribute results to a particular one), or are not clearly explicated, it is more difficult to say definitively that this particular class of materials has a particular pattern of effects. There are very tentative suggestions that the manner in which the woman is portrayed in the material (i.e., whether she is portrayed in a demeaning or degrading fashion) might be an important content factor but this is clearly an area that should be investigated. Certainly, the theoretical, (and many will argue the common-sensical) reasons for mediating effects on the basis of content cues are already available from social learning theory (Bandura, 1977; Bandura, et al., 1975).

Some Methodological Considerations

As we have done for previous sections describing different types of data collection procedures with different populations, we need to consider certain issues that pertain to experimental studies that will help clarify our evaluation of research findings. We will consider five issues in particular which are probably most often mentioned: the problem of the ability to generalize the results outside of the laboratory (what researchers call "external validity"); the problem of "the college student" as volunteer subject; the measures used to reflect "anti-social behavior"; ethical issues; and the operationalization of "pornography."

Ability to Generalize Experimental Findings

The problem of the "artificiality" of the experimental situation is an issue not new to social psychologists (see discussions by Berkowitz and Donnerstein, 1982; Littman, 1961). While it is true that the experiment is indeed "artificial," it is so by design. If one wanted to examine if X "causes" Y, a necessary condition for establishing such a casual connection is the elimination or control of other factors which may also affect Y. Such a condition then obviates a "real-world" setting in which numerous factors interact and jointly impinge on the individual. Littman (1961) maintains that systematic experimental designs are designed to test "more universal theoretical propositions that apply to large groups of human beings." That is, they are designed to test theorized relationships about human behavior that makes the issue of representativeness of the experimental setting and subjects of lesser consequence. Berkowitz and Donnerstein, 1982, offer a cogent summary of arguments on this point. (See also Kruglanski, 1975).

The College Student as Experimental Subject

The issue of representativeness has also been raised with regard to the college student as experimental subject, with the implication that the college student hardly represents "real people" in the "real world:" To reduce the issue to one of demographics is an oversimplification. If we are interested in the question of human response to sexually explicit materials, why should being in college or being male for that matter be a problem? As Berkowitz and Donnerstein (1982) point out, "The meaning the subjects assign to the situation they are in and the behavior they are carrying out plays a greater part in determining the generalizability of an experiment's outcome than does the sample's demographic representativeness or the setting's mundane realism." (p. 249)

Having said that, we also need to point out that there are, in fact, other attributes of the subject who participates in experiments involving exposure to sexually explicit materials that might have an impact on the interpretations of experimental results. Results from various studies suggest that:

1. Males, more than females, are likely to volunteer for sex-related experiments (Kendrick, et al., 1980).
2. Subjects who are willing to watch sexually explicit materials also tend to be sexually liberal, more sexually experienced, less anxious about sexual performance, and have fewer objections to pornography (Kaats and Davis, 1971; Farkas, et al., 1978; Wolchick Spencer and Lisi, 1983; Wolchick, Braver and Jensen, 1985).
3. Volunteer rates drop for both men and women the more intrusive the experimental conditions. Volunteer rates dropped by two-thirds (from thirty-eight percent to thirteen percent) for women and by over half for men (from sixty-seven percent to thirty percent) with the requirement of partial undressing to accommodate physiological arousal measurements (Wolchick, Braver and Jensen, 1985).

If participants are in fact more liberal, more experienced, and more accepting of sexually explicit materials, then it is certainly plausible that the "error;" if there is one, might be in the direction of null findings, while observed effects, particularly in the short term, might be indicative of their robustness (Eysenck, 1984). In any case, it is apparent that these other attributes ought to at least be considered in both the design and interpretation of experimental studies involving sexually explicit materials.

Ethical Considerations

While some bias may be inherent in the volunteer subject in general (Rosenthal and Rosnow, 1969) and in the volunteers for experiments involving sexually explicit materials in particular, we are constrained even more by understandable concerns regarding the more "vulnerable" segments of the population. Sherif's (1980) observations about the lack of evaluation procedures for the effectiveness of debriefing subjects in one particular study (see Malamuth, Heim and Feshbach, 1980 for the study in question and Malamuth, Feshbach, and Heim, 1980 for response to Sherif, 1980) have prompted researchers to measure debriefing effects (Malamuth and Check, 1984; Linz, 1985; Krafka, 1985) and also to eliminate from participation those who might be more vulnerable to the effects of exposure to materials in these studies. For example, Linz (1985) measured potential subjects on a psychoticism scale and eliminated from participation those who had high scores on this measure. Krafka (1985) excluded from her female subject pool those who were sexually inexperienced because of earlier findings (Wishnoff, 1978) that when these types of females were exposed to explicit erotic films, their sexual anxiety diminished while their expectations about engaging in sexual intercourse increased. The trade-off between ethical concerns and representativeness is evident in Krafka's observation: "Although this restricts the population to which the present results generalize, the author was unwilling to show sexually inexperienced females degrading images of sexual behavior and, especially, pornographic rape depictions." (p. 17)

These efforts to protect subjects from potential harm are, of course, laudable and a healthy response to concerns that have been raised. In terms of the final pool of subjects who participate in pornography experiments, however, the self-selection process described above and the researcher-imposed selection process must circumscribe our evaluation of research results.

Measures of Behavioral Effects

The range of dependent variable measures used in these studies is reasonably diverse. The use of similar measures across studies allows for better validation and the use of varied measures also provides the advantage of convergent validation. We will focus on behavioral measures of effects in this discussion and briefly discuss how attitude measures may or may not predict behavior.

Four categories of behavioral measures have been used in these studies:

1. Measures of aggressive behavior.

   The Buss aggression machine, sometimes known as a "shock box," has been widely used in laboratory experiements in the area of media violence and aggressive behavior (see reviews by Andison, 1978; Comstock, In Press). Donnerstein and his colleagues have used this measure to examine similar effects of exposure to violent and nonviolent pornography and aggressive behavior (Donnerstein, et al., 1975; Donnerstein and Barrett, 1978; Donnerstein, 1980; Donnerstein and Berkowitz, 1981).

   The procedure usually involves putting the subject in a "learning" situation where the subject's task as "teacher" is to make sure that a "learner" (usually an experimental confederate) masters a given lesson. When the learner makes a "correct" response, the subject is instructed to reward him or her by pressing a button illuminating a light. Whenever the learner makes an error, he is punished by means of an electric shock. The sequence of response has, of course, been preprogrammed. The subject's "aggressive tendencies" are recorded by means of the intensity and the duration of the shock which, in reality, is not received by the confederate (see Baron, 1977b for discussion on this measure).

   While this procedure has been criticized (see, for example, Baron and Eggleston, 1972), subsequent procedural modifications have increased its validity and has, in fact, been found to be highly predictive of physical aggression (Baron, 1977b). The question, however, of this measure's predictive validity in the area of sexually aggressive behavior outside of the laboratory still remains open since no efforts have been directed at examining this question.

   Other surrogate measures of aggressive behavior have included the infliction of aversive noise (Cantor, et al., 1978; Malamuth, 1983) and infliction of "pain" to an experimental confederate in a retaliation move where the subject has the opportunity to apply too much cuff pressure in a blood pressure reading situation (Zillmann and Bryant, 1984). Some validation is offered by Malamuth (1983) for the use of the aversive-noise measure with evidence that attitudes about realworld aggression (such as wife battering and rape) are clearly correlated with levels of laboratory aggression against females, suggesting some linkage between laboratory aggression and external responses outside the laboratory.

2. Judgments toward sexual assailants.

   In numerous studies, dependent measures have been obtained by having subjects respond to a rape case by evaluating both the victim and the assailant. While perceptual measures are most often used in this instance, one could also presumably consider delivering a verdict or a sentence as "behavior." In these instances, the presentation of a mock trial situation provides an element of mundane realism to the experimental situation. The studies by Linz (1985) and Krafka (1985) are excellent attempts at further diminishing demand characteristics of the experimental situation since the location of this phase of the experiment was conducted at the law school moot court where subjects were asked to evaluate what is purported to be the details of an actual rape case. An earlier study, a field experiment, by Malamuth and Check (1981) provides what may be the best procedure for eliminating demand characteristics and the measurement of effects in a setting that affords both control and realism. In this study, subjects were asked to watch the experimental films which were being shown on campus as part of the regular campus film program. Dependent measures were obtained a week later in what was presented as a public opinion survey. More studies in this area are clearly called for.

3. Choice and Viewing of Pornographic Fare.

   Zillmann and Bryant (In Press) utilized a unique way of measuring behavioral effects of exposure by examining subjects' choice of entertainment fare in an unobtrusively measured procedure. In their study of the effects of massive exposure, the following procedure was used to determine subjects' preferences for entertainment fare after they had been repeatedly exposed to pornography or to a neutral stimulus in the control condition: the subjects were met individually by the experimenter and informed of a brief delay caused by equipment problems. The subject was then taken to another waiting area (ostensibly another student's office) with a television set, a video tape recorder, and some video tape cassettes (including general interest and adult tapes ranging from "common erotica to graphic depictions of relatively uncommon sexual practices") and invited to feel free to watch. To ensure the subject knew he could watch in privacy, the subject was told the experimenter would call him on the phone to report to the designated room. Unknown to the subject was the fact that each cassette tape was programmed to emit a unique signal such that when the tape was played, an event recorder also recorded the amount of time spent watching.

   The advantage of this procedure is its experimental as well as ecological realism.

4. Self-reports of Aggressive Sexual Behavior.

   Two types of measures have been used to describe sexually aggressive behavior: a behavioral inclination measure operationalized by a self-reported likelihood of raping and using force in sexual interactions (see Malamuth, Haber and Feshbach, 1980; Malamuth, 1981; Briere and Malamuth, 1983) and a self-report inventory developed by Koss and Oros (1982) and used in several studies (see Malamuth, 1982; Malamuth, in press; Check, 1985). The latter includes a range of sexual behavior measures from saying things one does not mean to obtain sexual access to using various degrees of physical force.

An instrument developed by Burt to measure attitude (1980) has been used in a number of studies (Koss, 1986; Linz, 1985; Krafka, 1985; Malamuth and Check, 1981; Malamuth, 1981) to tap three dimensions: the acceptance of rape myths, the acceptance of interpersonal violence; and the acceptance of violence against women. The following are examples of the rape myth acceptance measure:

When women go around braless or wearing short skirts and tight tops, they are just asking for trouble.

Women who get raped while hitchhiking get what they deserve.

In evaluating these attitudinal measures and the laboratory measures of sexual behavior, two important questions have been raised to which we have alluded earlier. First, do attitudes predict behavior? And second, do laboratory measures of aggressive behavior predict actual aggression behavior?

On the first question, Malamuth and his colleagues have demonstrated a consistent correlation between Burt's (1980) attitudinal measures and their own measures of behavioral intentions (Briere and Malamuth, 1983; Malamuth, 1981; Malamuth, Haber and Feshbach, 1980; see also Malamuth and Briere, 1986 for a discussion on the attitude-behavior question in the area of sexual aggression). Koss (1986) has similarly demonstrated a high correlation between these sex-stereotyped beliefs and self-reports of sexual aggression. We do not have these same attitudinal data from those members of the population who provide the more extreme measures of sexually aggressive behavior-rapists-which might provide another means of validating the attitude-behavior postulate. However, interviews of incarcerated rapists appear to show similar acceptance of rape myths (Scully and Marolla, 1984). A number of studies are also reviewed in Malamuth and Briere (1986) which support the correlation between attitudes and non-laboratory aggressive behavior.

Operationalizations of Pornography.

Researchers, like lay people or the courts, have had some differences in the operationalization of "pornography." Malamuth (1984), for instance, uses the term with the qualifier that "no pejorative connotation is intended" and points out the difficulty of operationalizing the distinction between "aggressive versus positive types of pornography" (p. 29). However, he also relies on Steinem's (1980) separation of "acceptable erotica," in Malamuth's terms, emphasizing the notion of what Steinem called "shared pleasure," from "objectionable pornography," or what Steinem referred to as "sex in which there is clear force, or an unequal power" and describes stimuli in his research as using material belonging to the latter. Others have similarly used the term to refer only to sexually violent material and have used "erotica" to refer to nonviolent sexually explicit material (Abel, 1985). Still others on occasion simply use the term "erotica" and employ subclasses of aggressive and non-aggressive "erotica." (Donnerstein, 1983). Senn (1985) and Check (1985) have operationalized pornography to include both sexually violent and nonviolent but degrading categories and have classified all other sexually explicit portrayals as "erotica:"

In examining the types of stimuli used in these studies (Figure 1), it is clear that a wide diversity of research stimuli has been employed. These have ranged from partial nudity (Baron, 1979; Baron and Bell, 1977) to various levels of sexual activity, from "implied" to "explicit," covering a varied range of behaviors-masturbation, homosexual and heterosexual intercourse, oral-genital and oral-anal intercourse, fellatio, cunnilingus, bondage, and bestiality. Sources of materials have also run the gamut from so-called stag films to mainstream sexually explicit magazines, "adult" videos from the neighborhood video store, and even sex education films (Schmidt and Sigusch, 1970; see also the earlier description of stimulus materials used in 1970 experimental studies; Check, 1985). The 1970 Commission found the term "sexually explicit materials" to have greater utility.

Comparison among studies has become hampered by the differences in stimulus materials. A common classification system has been to make use of two subclasses: violent and nonviolent pornography (see Donnerstein, 1983, 1984) and while the stimulus materials representing the former have been relatively consistent (usually a rape scene with variations on victim reactions), the same cannot be said for "violent pornography." The full range of stimuli mentioned earlier, from partial nudity to bestiality (used, for instance, by Zillmann, Bryant, Comisky and Medoff, 1981) falls within the "nonaggressive" pornography category. Perhaps not surprisingly, a full range of results (negative, no effect, and positive) has also been elicited.

Donnerstein (1983) has maintained that differential arousal levels evoke different reactions, with "mild erotica" producing a pleasant distraction and more strongly arousing material resulting in negative effects. However, this differential-arousal attribute has not been pursued in subsequent studies. Zillmann and Sapolsky (1977) have suggested that in addition to arousal, the stimulus' valence property-how pleasing or displeasing it is-also accounts for differential findings.

If the effects from exposure to nonaggressive sexually explicit materials are mediated in part by their affect value, a problem still remains: how do we explain the "pleasing" or "displeasing" character of a stimulus? Pleasing or displeasing evaluations could arise from a number of factors including the explicitness of the material, the type of activity portrayed (see, for example, Glass' [1978] scale analysis of the 1970 Commission survey data which shows clearly gradations in public perceptions of different activities), or the theme employed. For example, Sherif (1980) raised the possibility of power differentials to explain female subjects' arousal but high negative affect in response to a stimulus portraying a rape victim experiencing an involuntary orgasm in Malamuth, Heim and Feshbach's (1980) study.

Two studies (Check, 1985; Senn 1985) have attempted to reconceptualize nonaggressive sexually explicit materials into two further classes ('sexually explicit and degrading or dehumanizing', and simply 'sexually explicit'). There is theoretical justification for expecting differential effects from these subclasses. Bandura, Underwood and Fromson (1975) have demonstrated that socially reprehensible attitudes or behaviors may be made more acceptable by dehumanization of victims. "Inflicting harm on individuals who are subhuman and debased is less apt to arouse self-reproof than if they are seen as human beings with dignifying qualities." (p. 255). Again, this is clearly a line of research that merits further attention.

The problem of explicating stimulus attributes is complicated with examination of a class of materials categorized by their commercial label: "R-rated slasher films" (see Linz, Donnerstein and Penrod, 1984; Linz, 1985; Krafka, 1985), or "X-rated films." The former "contain explicit scenes of violence in which the victims are nearly always female. While the films often juxtapose a violent scene with a sensual or erotic scene (e.g., a woman masturbating in the bath is suddenly and brutally attacked), there is no indication in any of the films that the victim enjoys or is sexually aroused by violence. In nearly all cases, the scene ends in the death of the victim." (Linz, et al., 1984, p. 137). These studies using this film genre have generally found desensitizing effects among male subjects, after massive exposure.

But the question still remains: what does this class called "R-rated slasher films" mean conceptually? If one were interested in describing potential effects from classes of sexually explicit materials, where does this set of materials fit in? This appears compounded in an examination of effects of sexually violent, violent, and sexually explicit materials on female subjects (Krafka, 1985), where these films are used to operationalize "violent" films, despite allowing that they have "some sexual content."

"X-rated films" pose the same problems. While they appear to be used to represent sexually explicit material without any violence, different themes may be emphasized leading to quite different results.

The need to utilize meaningful classes that go beyond those in current use is important not just for validity requirements. After all, the question which social scientists must ultimately address-with both theoretical and pragmatic or public policy implications-is what types of effects have been demonstrated for what classes of material? Such investigations for some social scientists may have undesirable political or idealogical implications but ignoring the issue also hampers our ability to explain the nature of effects more fully so as to provide for nonlegal policy strategies that are firmly anchored in social science findings (see, for example, Byrne and Kelley, 1984; Kelley, 1985).

Some Theoretical Considerations

In designing research studies to answer particular questions, social scientists do not ordinarily operate in a vacuum. Quite often, the relationships posited, the selection of variables and their operationalizations, the groups of people selected for examination, and the general research procedures are guided by "theory." Quite simply, this is the explanatory framework which rationalizes or justifies why a particular relationship might be expected.

We think it useful to summarize some of the theoretical reasoning that has been applied to the general question of what effects if any might be found from exposure to sexually explicit stimuli.

1. Social Learning Theory

   This approach offers a perspective on human behavior based on the notion that there is "a continuous reciprocal interaction" between environmental factors, an individual's processing of information from his environment and his behavior (Bandura, 1977). This framework assigns a prominent role to the processes of vicarious and symbolic learning (i.e., learning by observing others' behavior and one's own) and a self-regulating process whereby an individual selectively organizes and processes stimuli and regulates his or her behaviors accordingly.

   The generic process of modeling is a major component of social learning which many mistakenly interpret as simply imitation, or a one-to-one correspondence between some portrayed novel behavior and the reproduction of such behavior. While this type of effect is not precluded (and there are certainly many anecdotal media accounts of such instances), "modeling" embraces a more complex array of processes which can be subsumed under two categories. First, modeling includes the facilitation of particular response categories ("response facilitation") which assumes that a portrayed behavior functions as an external inducement for similar sets of responses which can be performed with little difficulty. Second, it includes the capacity to strengthen or weaken inhibitions of responses ("inhibition" or "disinhibition") that may already be in the observer's repertoire. If there are restraints on a particular behavior (self-restraints, as in anxiety over a particular behavior, or external restraints, including the possibility of getting caught and punished for some socially disapproved-or illegal-action), such restraints may be lifted when an observer sees a model engage in disapproved acts without any adverse consequences (Bandura, 1973, 1977).

   In Check and Malamuth's (1985) application of this theoretical framework, they discuss their findings in terms of Bandura's postulated "antecedent" and "consequent" determinants. The former incorporates symbolic expectancy learning principles exemplified by the symbolic pairing of sex with violence against women and vicarious expectancy learning, or observing others becoming aroused to sexual violence. Consequent determinants include observing seeing a male use force, not getting punished, and, furthermore, finding the experience pleasurable for himself and for his victim.

   Two studies based on survey data provide additional information that certain sexually explicit materials may provide "models" for behavior for some individuals.

   Russell (1985) reported findings from an earlier study on sexual abuse of women. A probability sample of 930 adult female residents in San Francisco were interviewed. Of this number, about four in ten (389 women) said they had seen pornography and forty-four percent of this group reported being upset by it. Fourteen percent of the total sample reported they had been asked to pose for pornographic pictures and ten percent said they had been upset by someone trying to get them to enact what had been seen in the pornographic pictures, movies or books. An additional finding in this study was that those who were upset by pornographic requests were twice as likely to be incest victims than those who were not upset by similar requests. A similar pattern was found among those who reported being upset at being asked to pose for pornographic pictures, i.e., those who were asked to pose were more than twice as likely to suffer incest abuse in their childhood (thirty-two percent versus fourteen percent). What this suggests, according to Russell, is that women who suffered sexual abuse are significantly more vulnerable to pornography-related victimization, a "revictimization" syndrome.

   Silbert and Pines (1986), in a similar study on sexual assault of street prostitutes, came upon unexpected information in the course of their interviews. From detailed descriptions the subjects provided to open-ended questions in regard to incidents of juvenile sexual assault in their childhood and to incidents of rape following entrance into prostitution, it became evident that violent pornography played a significant role in the sexual abuse of street prostitutes. Of the 200 prostitutes interviewed, 193 reported rape incidents and of this number, twenty-four percent mentioned allusions to pornographic material on the part of the rapist. Since these comments were not solicited, it is likely that this figure is a conservative estimate. The authors described the comments as following the same pattern: "the assailant referred to pornographic materials he had seen or read and then insisted that the victims not only enjoyed the rape but also the extreme violence." (p. 12)

2. Arousal

   Arousal has been conceived of as a "drive" that "energizes or intensifies behavior that receives direction by independent means" (Zillmann, 1982, 1978). This model relies on the notion that arousal based on exposure to some communication stimulus can facilitate behaviors which could either be prosocial or anti-social, depending on situational circumstances. Such circumstances could include specific content cues which might elicit either positive or negative affect (Sapolsky, 1984). If arousal levels are minimal and the stimulus evokes pleasant responses (as might be the case when viewing mildly erotic material), the effect might be reduced aggression. If, on the other hand, the stimulus elevates arousal to high levels, then the outcome might be aggressive behavior. This approach has been criticized for its inability to account for the predominance of one response rather than another.

3. Habituation

   The idea of habituation is akin to drug treatment or drug dependency where, over time, one must rely on increasing doses to obtain the same effect. In the area of exposure to explicit sexual stimuli, repeated exposure has resulted in initially strong arousal reactions becoming weaker over time, leading to habituation. (Zillmann, 1982, 1984). One attitudinal manifestation of this effect is callousness, either to victims of aggression or simply to the violent or anti-social behaviors themselves. While this holds promise as an explanatory framework, more research is needed, particularly longitudinal studies, to demonstrate its predictive utility.

4. Cue Elicitation/Disinhibition

   Berkowitz (1974,1984) has proposed a stimulus-response relational model which suggests that an individual (e.g., a film viewer) reacts impulsively to environmental stimuli and this reaction is determined in part by predispositions and in part by stimulus situational characteristics which could function to "disinhibit" such predispositions. Berkowitz has demonstrated that cues associated with aggressive responding such as a situation depicting a female victim, when viewed by an individual predisposed to aggress (one who is provoked or angered), will more likely evoke the aggressive response as a result of the stimulus-response connection already established by previous exposure to the films. (See Donnerstein and Berkowitz, 1981, and Linz, 1985, for applications).

These explanatory-predictive approaches may not necessarily operate independently; they could conceivably complement each other. They stand, however, in contrast and direct opposition to the catharsis theory which is still being promoted in many quarters as the explanation for why exposure to sexually explicit materials has only beneficial effects. Catharsis suggests that exposure to highly arousing material actually leads to a diminution of anti-social effects because relieving the arousal then reduces the instigation to commit any sex crimes in the future. Unfortunately, little evidence exists for this claim and numerous research reviews (primarily in the area of media violence and aggressive behavior) have arrived at this same conclusion (Berkowitz, 1962; Bramel, 1969; Weiss, 1969; Geen and Quanty, 1977; National Institute of Mental Health, 1982; Comstock, 1985). The following observation typifies comments made about the catharsis theory.

The cause-effect hypothesis that we already described is not supported by the data. Little evidence for catharsis, as we have defined it, exists and much of the evidence that has been adduced in its favor is susceptible to alternative explanations that are at least parsimonious. In fact, when conditions that give rise to such alternative explanations are removed from the experimental setting, the reverse [authors' emphasis] of what the catharsis hypothesis predicts is usually found, i.e., aggression begets more, not less, aggression (p. 6).

It is instructive that some have called a moratorium on catharsis (Bandura, 1973), others have proclaimed its demise (Comstock, 1985). Even its major proponent has reformulated his position by explaining why it does not apply to situations involving media exposure (Feshbach, 1980).

Other Effects of Sexually Explicit Materials

If we take the entire potential range of "effects" which could occur as a result of exposure to sexually explicit materials, and if we take the commission of sex offenses to be one extreme of that continuum, then the other end might be represented by beneficial effects. Many have made an argument for such benefits (Tripp, 1985; Wilson, 1978).

Public opinion data both in 1970 and in 1985 show that a majority believe use of sexually explicit materials "provide entertainment," relieve people of the impulse to commit crimes, and improve marital relations.

If they are any indication, the popularity of "How-To" articles on sex in the popular media and in best-sellers such as The Joy of Sex, The Sensuous Woman, and others like them are also testament to the learning that might occur from these materials.

There are also two areas in which sexually explicit materials have been used for positive ends: the treatment of sexual dysfunctions and the diagnosis and treatment of some paraphilias.

In the area of sexual dysfunctions, a common conceptual model views a particular goal as a new response to be learned. The reduction of sexual anxieties or the attainment of orgasm for nonorgasmic individuals might be examples of such objectives. In the process of learning a new response, two steps are implicated: the weakening of response inhibitions and facilitation of the acquisition of new behavior patterns that comprise the steps toward the final objective.

For instance, in teaching nonorgasmic females to achieve orgasm, therapeutic procedures might include desensitization techniques, followed by the modeling of a hierarchy of behaviors such as body exploration, genital manipulation, self-stimulation to orgasm, and the generalization of the response to a partner (Caird and Wincze, 1977; LoPiccolo and Lobitz, 1972; Heiman, LoPiccolo and LoPiccolo, 1976).

A number of controlled experimental studies have demonstrated the efficacy of therapeutic treatments involving video taped modeling, written instructions which implicate principles of observational learning, and information processing. Such procedures have been successful in changing both attitudes and behaviors (Anderson, 1983; Heiby and Becker, 1980; Nemetz, Craig and Reith, 1978; Wincze and Caird, 1976; Wish, 1975).

In the case of diagnosis and treatment of sex offenders, the identification of arousal patterns and the subsequent therapy program (which might involve the inhibition of inappropriate arousal responses such as arousal to a photograph of a child) have involved the use of sexually explicit materials. As part of some treatment methods, the use of aversive techniques might be directed at extinguishing deviant arousal, or they might be combined with positive reinforcement for more appropriate sexual responses. In some treatment programs, the combination of these procedures with social skills training has been found to be effective (Abel, Becker and Skinner, 1985; Whitman and Quinsey, 1981). However, the results have been less conclusive for narrower approaches to treatment (see Quinsey and Marshall, 1983).

On the whole, the learning principles that include vicarious learning, reinforcement, disinhibition principles that are used in these therapeutic controlled settings are no different from those which have been employed to explain the acquisition of negative attitudes and behaviors.

An Integration of the Research Findings

It is clear that the conclusion of "no negative effects" advanced by the 1970 Commission is no longer tenable. It is also clear that catharsis, as an explanatory model for the impact of pornography, is simply unwarranted by evidence in this area, nor has catharsis fared well in the general area of mass media effects and anti-social behavior.

This is not to say, however, that the evidence as a whole is comprehensive enough or definitive enough. While we have learned much more since 1970, even more areas remain to be explored.

What do we know at this point?

• It is clear that many sexually explicit materials, particularly of the commercial variety, that are obviously designed to be arousing, are, in fact, arousing, both to offenders and nonoffenders.
• Rapists appear to be aroused by both forced as well as consenting sex depictions while nonoffenders (our college males) are less aroused by depictions of sexual aggression. On the other hand, when these portrayals show the victim as "enjoying" the rape, these portrayals similarly elicit high arousal levels.
• Arousal to rape depictions appears to correlate with attitudes of acceptance of rape myths and sexual violence and both these measures likewise correlate with laboratory observed aggressive behaviors.
• Depictions of sexual violence also increase the likelihood that rape myths are accepted and sexual violence toward women condoned. Such attitudes have further been found to be correlated with laboratory aggression toward women. Finally, there is also some evidence that laboratory aggression toward women correlates with self-reported sexually aggressive behaviors.
• What we know about the effects of nonviolent sexually explicit material is less clear. There are tentative indications that negative effects in the areas of attitudes might also occur, particularly from massive exposure. The mechanics of such effects need to be elaborated more fully, however, particularly in light of more recent findings that suggest that degrading themes might have effects that differ from non violent, non degrading sexually explicit materials. This is clearly an area that deserves further investigation.
• There are suggestions that pornography availability may be one of a nexus of sociocultural factors that has some bearing on rape rates in this country. Other cross-cultural data, however, offer mixed results as well, so these findings have to be viewed as tentative at best.
• We still know very little about the causes of deviancy and it is important to examine the developmental patterns of offenders, particularly patterns of early exposure. We do have some convergence on the data from some rapists and males in the general population in the areas of arousal and attitudes but again, this remains to be examined more closely.

Clearly, the need for more research remains as compelling as ever. The need for more research to also examine the efficacy of strategies for dealing with various effects is as compelling. If learning-both prosocial and antisocial-occurs from various depictions, and there certainly is clear evidence of both, the need for strategies that implicate the same learning principles must be evaluated. Educational and media strategies have been discussed elsewhere and found to be effective in such disparate areas as health and media violence (see Rubinstein and Brown, 1986; Johnston and Ettema, 1982; American Psychological Association, 1985). Researchers in the area of pornography have no less a responsibility.

Summary of Commission Findings of Harm from Pornography

The Commission divided pornography into four classifications and then analyzed each classification according to three tiers as set forth below:

1. Sexually Violent Materials
   1. Social Science Evidence. -- Negative effects were found to have been demonstrated.

   2. Totality of Evidence. -- Harm found in all sub-tiers:
      1. Acceptance of Rape Myths;
      2. Degradation of the Class/Status of Women;
      3. Modeling Effect;
      4. Family
      5. Society
   3. Moral, Ethical and Cultural. Harm found.

2. Sexual Activity Without Violence But with Degradation, Submission, Domination, or Humiliation
   1. Social Science Evidence. -- Negative effects were found to have been demonstrated.

   2. Totality of Evidence. -- Harm found in all subtiers:
      1. Acceptance of Rape Myths;
      2. Degradation of the Class/Status of Women;
      3. Modeling Effect;
      4. Family
      5. Society
   3. Moral, Ethical and Cultural. -- Harm found.

3. Sexual Activity Without Violence, Degradation, Submission, Domination, or Humiliation

   All Commissioners agreed that some materials in this classification may be harmful, some Commissioners agreed that not all materials in this classification are not harmful. It was determined that this classification is a very small percentage of the total universe of pornographic materials. See text for further discussion.

4. Nudity Without Force, Coercion, Sexual Activity, or Degradation

   All Commissioners agreed that some materials in this classification may be harmful, some Commissioners agreed that not all materials in this classification are not harmful. See text for further discussion.

Notes

1119. Surveys such as this Gallup survey which employ "probability samples" are generally accurate within known limits. That is, the sample results can be applied to the population as a whole within the sampling tolerance ranges for a given sample size. For this survey sample size of 1020 respondents, sampling error is three percent. In practical terms, if we could contact every member of the population being described, the "real" percentage would be within plus or minus three percent of the observed percentage for the sample.

1120. The Washington Post-ABC News poll was conducted February 6-12, 1986 by telephone among 1,504 men and women nationwide. The margin of error is plus or minus three percentage points.

1121. The READEX survey was part of the public record as part of a court case involving Hustler magazine (Herceg v. Hustler, Inc.).
1122. Baron and Straus recently conducted additional analyses of their data by introducing a "Violence Approval Index;" based on attitude measures from the general Social Survey. By introducing this into their original equation, the relationship between the sale of sex magazines and rape disappeared. While this could offer some tentative support for the authors' contention that a "hypermasculine" climate might be responsible for rape rates, rather than sex magazines per se, they are also appropriately cautious about the severe limitations of this particular finding. They point out that while the Violence Approval index correlates in expected fashion with the percent males in the population, the percent of the population in the high-violence age group of eighteen to twenty-four, with the Legitimate Violence and Social Disorganization indices, it also has an unexpected negative correlation with the percent single males in the population and has a low correlation(.23) with the rape rate. Second, the data are restricted to forty states which, in combination with the addition of still another variable to the equation, increases the standard error. Until these problems are sorted out, the impact of this variable will have to remain speculative. It is presumably for this reason that these authors included this information in a footnote rather than in their text, and we likewise do so.
1123. The question used by Koss (1986) in this survey as a measure of pornography exposure was:

How often do you read any of the following magazines: Playboy, Penthouse, Chic, Club, Forum, Gallery, Genesis, Oui, or Hustler? (check one):

	________ Never
	________ Seldom
	________ Somewhat frequently
	________ Very frequently

1124. The following films were used: R-rated nonviolent "teen sex" films-Porky's, Fast Times at Ridgemont High, Private Lessons, Last American Virgin, and Hots. X-rated nonviolent films-Debbie Does Dallas, Health Spa, The Other Side of Julie, Indecent Exposure, and Fantasy. R-rated "slasher" films; Texas Chainsaw Massacre, Maniac, Toolbox Murders, Vice Squad, and I Spit on Your Grave.
1125. Psychoticism measures included such items as the following (Linz, 1985):
      • The idea that someone else can control your thoughts.
      • Having thoughts about sex that bother you a lot.
      • The idea that something is seriously wrong with your body.
      • Never feeling close to another person.
      • Feeling lonely when you are with other people.

[www.porn-report.com] Attorney General's Commission on Pornography (1986)   Last| Home | Next

Layout and design
(c)2004 www.porn-report.com