Attitudes Toward Nomarital Sex in 24 Countries.

Comparing sexual attitudes and behaviors across cultures is a concern of anthropological and sociological research. Scholars have pointed out that many societies around the world share common normative attitudes toward sex, including the incest taboo, condemnation of adultery, and a general concern for regulating sexuality, particularly outside wedlock (Levi-Strauss, 1969; Murdock, 1960). At the same time, research has shown that attitudes toward premarital sex, homosexuality, masturbation, petting, and other sexual behavior are not constants, but present different patterns from culture to culture (Ford & Beach, 1951). Even in the Western world, countries differ in their emphasis on the value of virginity before marriage, the acceptability of homosexuality, and the appropriate age to become sexually active (Buss, 1989; Christensen & Carpenter, 1962; Christensen & Gregg, 1970; Jones et al., 1985; Jones et al., 1986; Ross, 1985).

Although scholars, policy makers, and the general public have a genuine interest in cross-national variations in attitudes regarding sexual morality, most research in the social and behavioral sciences has been largely a local or national enterprise. Indeed, most studies have been based on samples not representative of a national population, and nationally representative studies have seldom been designed with an eye to facilitating international comparisons. Cross-national research on sexuality remains relatively underdeveloped.

Considering nonmarital sex, attitudes toward premarital sex, teenage sex, extramarital sex, and homosexual sex are sometimes assumed to be indicators of a single underlying dimension measuring the overall sexual permissiveness of individuals or nations. While these four variables are correlated, it need not follow that societies show consistently permissive or conservative attitudes on all four variables, nor that a simple permissive-nonpermissive dichotomy is enough to account for variation across countries. If moral evaluations of different sorts of nonmarital sexual relationships hang together in complex ways, this implies a more pluralistic landscape of sexual values than a permissive nonpermissive dichotomy would suggest. For example, nations which are tolerant of sexual experimentation before marriage might be more approving of both premarital sex and teen sex, but not necessarily of extramarital sex. Nations where many people are critical of marriage as a social institution might support the right of premarital cohabitors and homosexuals to have sex, but not necessarily young teens. We therefore hypothesize that these four variables combine in different ways across countries, and that countries form clusters distinguished by characteristic profiles of attitudes toward nonmarital sex.

How many different sexual regimes exist and which countries belong to what regime are empirical questions. These questions are addressed using newly available national survey data from 24 countries on attitudes toward homosexual sex, extramarital sex, premarital sex, and teenage sex. Moving beyond previous research, we are interested in measuring how much of attitudes toward nonmarital sex is shared by countries or regimes. Researchers, particularly anthropologists, have been interested in the extent to which cultures share sexual norms, values, and ethics, as part of a broader interest in universal features of social life that transcend particular cultures and idioms. Even though we are mostly concerned with modern industrial societies, we also wish to address the issue of universality versus specificity, and we introduce a straightforward way of measuring the degree of sharing.

Research on Cross-National Differences

Research on sexual attitudes in industrialized nations points out the fact that normative standards differ across countries. For example, comparative research has repeatedly shown that Scandinavian students are more tolerant of premarital sex than are their American counterparts (Christensen & Gregg, 1970; Schwartz, 1993; Weinberg, Lottes, & Shaner, 1995). A 37-country study reported that non-Western societies (China, India, Indonesia, Iran, Taiwan, and Palestinian Arabs) value chastity highly in a potential mate, that Western European countries (Sweden, Norway, Finland, the Netherlands, West Germany, and France) attach little importance to prior sexual experience, and that Ireland, Japan, and Poland fall somewhere in between (Buss, 1989). Data on attitudes toward homosexuality have been reported for 16 industrialized countries. They suggest that tolerance for homosexuality is linked with postmaterialist values (Inglehart, 1990). Tolerance for sexual expression in the public sphere has also been shown to differ markedly between Western, industrialized nations: Sweden and Ireland anchor the extremes in terms of openness in the sale of sexually explicit material, public nudity, and condom advertising (Jones et al., 1985).

Despite the cumulation of comparative findings, cross-national research has been limited in several ways. First, although anthropologists have undertaken cross-cultural analyses of a broad range of sexual attitudes and behaviors (Broude & Greene, 1976; Ford & Beach, 1951; Minturn, Grosse, & Haider, 1969; Murdock, 1960), cross-national research on complex, industrial societies has focused on only a few aspects of sexuality. Premarital sexuality, including teenage sex, has been emphasized, largely to the exclusion of other kinds of nonmarital sexual relationships. Little is known about how attitudes toward extramarital sex and homosexual sex vary across countries.

Second, comparative research on complex, industrial societies has focused on relatively few countries. Most articles that we have reviewed concern the United States and a Northern European country. Sweden, for example, has been posed as a counterpoint to American views of sexual morality. Relatively little comparative information is available on other European countries and even less on non-European countries. Comparing the same handful of countries again and again dooms research to redundant findings. Even our understanding of well-studied nations is advanced by contextualizing them within a broader set of contemporary societies.

Third, most existing research dealing with international comparisons rests on samples which are not nationally representative. Comparative studies tend to make good use of small convenience samples of college students or other groups strategically chosen to elucidate a particular issue (Christensen & Carpenter, 1962; Kyes & Tumbelaka, 1994; Lottes & Weinherg, 1997; Ma, 1989; Perlman, Josephson, Hwang, Begum, & Thomas, 1978; Ross, 1985; Schwartz, 1993; Sprecher & Hatfield, 1996; Sprecher, Hatfield, Cortese, Potapava & Levitskaya, 1994; Weinherg et al., 1995). Buss's (1989) 37-country study, for example, was not based on probability sampling but relied instead on various sampling strategies, including systematic household sampling in Venezuelan neighborhoods, mail responses to West German newspaper advertisements, and Estonian couples applying for marriage licenses.

Fourth, even though the representativeness of data on sexuality has been advanced by some excellent national surveys,including the 1992 National Health and Social Life Survey in the United States (NHSLS) (Laumann, Gagnon, Michaels, & Michaels, 1994), the 1990-91 British National Survey of Sexual Attitudes and Lifestyles (Wellings, Field, Johnson, & Wadsworth, 1994), and the 1992 Analysis of Sexual Behavior in France Survey (ACSF) (Bozon & Leridon, 1996), there is a dearth of genuinely comparable studies. Differences in question wording, response categories, survey context, and procedures undermine confidence in comparability. An exemplar for the standardized collection of information relevant to AIDS has been developed (Carballo, Cleland, Carael, & Albrecht, 1989), leading to a report of cross-national sexual attitudes and behaviors in 16 mostly developing countries (Cleland & Ferry, 1995). Typically, researchers who wish to understand cultural differences and similarities have had to overlook problems of strict comparability. For example, to demonstrate that men report more sexual partners than do women, Smith (1992a) relied on American, Canadian, British, and Norwegian surveys which, while nationally representative, differed in question wording, response categories, and time referents. The cross-national World Values Surveys contain comparable data on whether extramarital sex, homosexual sex, and sex under the legal age of consent are ever justified, but these items have not been exploited to examine regimes of attitudes toward nonmarital sex (Inglehart, 1997).

DATA

In contrast to previous research, this study has the benefit of data from a standardized questionnaire collected in 24 countries from large and nationally representative samples. Analysis capitalizes on data from the International Social Survey Program (ISSP) made available by the Zentralarchiv Fuer Empirische Sozialforschung, Koeln(1) The ISSP is an established program of cross-national collaboration which, since 1985, has facilitated annual social science surveys (Smith, 1992b). In 1994, independent research institutions in individual countries replicated survey questions on family, sexuality, and gender roles, typically as a supplement to national surveys. Data are available for 24 countries, including Australia, Germany (East and West reported separately), Great Britain, Northern Ireland, the United States, Austria, Hungary, Italy, Ireland, the Netherlands, Norway, Sweden, Spain, the Czech Republic, Slovenia, Poland, Bulgaria, Russia, New Zealand, Canada, the Philippines, Israel, and Japan. The total sample includes 33,590 respondents. National samples range in size from 647 in Northern Ireland to 2,494 in Spain. Appendix A presents characteristics of the samples and their design by country.

Survey informants were asked about four types of nonmarital sex. The premarital sex item read "Do you think it is wrong or not wrong for a man and woman to have sexual relations before marriage?" This was followed by questions asking "What if they are in their early teens, say under 16 years old?"; "What about a married person having sexual relations with someone other than his or her husband or wife?"; and "What about sexual relations between two adults of the same sex?" The respondents were presented with four response categories: always wrong, almost always wrong, wrong only sometimes, or not wrong at all. Other responses (it depends, cannot choose, don't know) were treated as missing values.

Wording of the questionnaire items was standardized by translating from British English, hut comparability may still have been affected by country differences in mode of data collection, sample design and population coverage, country-specific content of the various questionnaires to which these items were appended, and other factors. Even allowing for possible measurement error and the limited number of items asked about sexuality, the ISSP data offer a considerable improvement over prior cross-national efforts because they present nationally representative data on attitudes toward four different types of nonmarital sex in 24 countries. The data are superior to those previously available to scholars interested in comparative analysis of sexual attitudes in contemporary industrialized societies.

RESULTS

Attitude Profile by Country

We first aggregated the 33,590 responses using percentages into a matrix of 24 rows (the countries) by 16 columns (the four response categories for each of the four types of nonmarital sex). This matrix appears in Table 1.

Table 1. Attitudes Towards Nonmarital Sex by Country (Percentages)

                       Sex before marriage wrong?

Countries                   Almost     Only      Not Wrong
                   Always   always   sometimes    at all

Australia            13       8         19         59
Austria               4       8         19         69
Bulgaria             23       9         11         57
Canada               12       5         15         69
Czech Republic        5       6         20         68
Germany (East)        2       1         10         87
Germany (West)       12       3         14         79
Great Britain        19       5         14         70
Hungary              19      10         17         54
Ireland              35       8         15         42
Israel               19       7          9         65
Italy                19       7         15         59
Japan                19      21         44         15
Netherland            7       3         13         77
New Zealand          19       6         14         62
Northern Ireland     31       8         13         48
Norway                7       3         13         77
Philippines          60      16         13         11
Poland               18       9         13         60
Russia               13      12         18         57
Slovenia              3       3         12         82
Spain                20       8          9         63
Sweden                4       2          5         89
USA                  29      12         18         41
Overall              17       7         15         61

                             Sex before 16 wrong?

Countries          Always   Almost      Only     Not wrong
                            Always   sometimes    at all

Australia            61      20         10          9
Austria              36      38         18          8
Bulgaria             71      14          7          8
Canada               55      20         16          9
Czech Republic       59      21         12          8
Germany (East)       27      22         34         16
Germany (West)       34      24         29         13
Great Britain        67      21          9          3
Hungary              50      26         17          7
Ireland              84       9          6          1
Israel               67      14          9          9
Italy                58      19         14          9
Japan                60      30          7          3
Netherland           45      24         25          5
New Zealand          71      18          8          4
Northern Ireland     81      12          5          2
Norway               55      24         15          6
Philippines          77      16          5          3
Poland               77      14          4          4
Russia               45      30         14         11
Slovenia             44      27         23          6
Spain                59      18          9         14
Sweden               32      28         22         17
USA                  71      16          9          4
Overall              58      21         14          7

                         Extramarital sex wrong?

Countries          Always   Almost      Only     Not wrong
                            Always   sometimes    at all

Australia            59      24         10          7
Austria              67      28          3          2
Bulgaria             51      20         14         16
Canada               68      19         10          2
Czech Republic       43      26         21          9
Germany (East)       60      23         13          4
Germany (West)       55      27         13          4
Great Britain        67      23          8          2
Hungary              62      20         15          3
Ireland              80      13          5          2
Israel               73      17          6          4
Italy                67      13         15          5
Japan                58      33          7          2
Netherland           63      24         11          2
New Zealand          75      17          6          1
Northern Ireland     81      12          6          2
Norway               70      25          5          1
Philippines          88      10          1          1
Poland               74      14          7          5
Russia               36      26         21         17
Slovenia             57      27         13          3
Spain                76      13          7          5
Sweden               68      26          5          1
USA                  80      14          4          2
Overall              66      21          9          4

                          Homosexual sex wrong?

Countries          Always   Almost      Only     Not wrong
                            Always   sometimes    at all

Australia            55       8         10         27
Austria              52      18         15         15
Bulgaria             81       5          4         10
Canada               39       5         10         46
Czech Republic       29      12         21         39

Germany (East)       51       9         10         30
Germany (West)       42      10         14         33
Great Britain        58       6         10         26
Hungary              83       8          5          4
Ireland              71       6          6         17
Israel               57       8          7         27
Italy                67       7          7         19
Japan                65      22         11          2
Netherland           19       4         12         65
New Zealand          56       5          8         31
Northern Ireland     80       4          6         10
Norway               47       6         11         37
Philippines          84       9          3          3
Poland               77       6          4         14
Russia               57      17          7         19
Slovenia             70      13          9          8
Spain                45       7          6         42
Sweden               56       6          6         32
USA                  70       4          7         19
Overall              59       9          9         24

Before considering whether countries cluster into distinctive sexual regimes, it is useful to examine the country-by-country distribution of attitudes appearing in Table 1. In each country, premarital sex is the most accepted of the four types of nonmarital sex-a finding consistent with cross-cultural research on indigenous societies (Brown, 1952). The mean percent agreeing that premarital sex is not wrong at all is 61% across the 24 countries. A substantial majority of respondents in most countries chose this response, but the exceptions are worth pointing out. For example, only 11% of Filipinos consider premarital sex to be not wrong at all. Japan also presents a unique pattern of responses: 44% of Japanese respondents indicate premarital sex is only sometimes wrong, as compared with a mean of 15% across all countries. The remaining countries reflect two general patterns: Either they demonstrate a strong consensus on the acceptability of premarital sex or, less commonly, they are polarized between strong approval and strong disapproval, as in the cases of Ireland, Northern Ireland, and the United States.

Approval of premarital sex is limited to adults, however. Although a majority see nothing wrong with sex before marriage, only 7% of respondents across countries voice this opinion when teenagers younger than 16 are concerned. In fact, 58% condemn young teenagers having sex as always wrong. Germany (East and West), Austria, and Sweden are exceptions which are much more tolerant than other countries. At the other extreme, Bulgaria, New Zealand, Ireland, Northern Ireland, the Philippines, Poland, and the United States present very low acceptance for teenage sex. Extramarital sex elicits, if anything, even stronger disapproval. Across 24 nations, only 4% report that it is not wrong at all for married people to have sex with someone besides their husband or wife. Despite a general consensus across countries that extramarital sex should be condemned, several nations, notably Russia, Bulgaria, and the Czech Republic, show considerably more tolerance. For example, only 36% of Russians are prepared to describe extramarital sex as always wrong, as compared to an average of 66% for all 24 countries.

While not as widely accepted as premarital sex, sexual relations between adults of the same sex do not prompt the strong disapproval that extramarital sex or teen sex does. On average, 24% of national populations regard homosexual behavior as not wrong at all as compared to 59% who view it as always wrong. Attitudes towards homosexual sex, however, show considerable variation from country to country. In most countries (18 of 24), a majority of people report homosexual sex as being always wrong. In the Netherlands, however, about two thirds of the population sees nothing wrong with sexual relations between two adults of the same sex. In Spain, the Czech Republic, Canada, and Norway substantial percentages voice similarly tolerant attitudes, although public opinion is polarized on the issue in those countries. Also accepting of homosexuality, albeit to a lesser extent, are Germany (East and West) and Sweden.

To sum up, with a few notable exceptions there is widespread acceptance of premarital sex across abroad sample of developed nations. This acceptance, however, does not extend to premarital sex between young teenagers, nor to extramarital sex. While more accepted than either extramarital sex or teen sex, homosexuality generates more varied opinions across countries and more polarized responses within nations.

Clusters of Countries

On a variable-try-variable basis, we were able to identify countries which are more or less alike in attitude towards a particular type of nonmarital sex. It remains to be seen whether countries can be typified according to distinctive sexual regimes; that is, characteristic attitude profiles with respect to all four nonmarital behaviors.

Given 24 countries, there could conceivably be anywhere from one to 24 regimes. We hypothesized, however, that a two-group solution, consistent with a simple permissive-nonpermissive dichotomy, would be insufficient to explain differences between countries. To find out whether countries can be gathered in significant clusters with specific profiles, we applied cluster analysis to the matrix in Table 1. Cluster analysis has been employed previously to elucidate differences between individuals in sexual attitudes (Laumann et al., 1994; Lottes, 1985). Although clustering methodology is often used in cross-national research, this paper is, so far as we can determine, the first to apply cluster analysis to sexual attitudes when contemporary industrial societies are the units of analysis.

To determine the number of regimes, we examined a sequence of hierarchical cluster analyses based upon Ward's method of clustering applied to the nation profiles reported in Table 1. Instead of partitioning the countries into some predetermined number of clusters in a single step, this hierarchical procedure produces step-by-step splits (Everitt, 1993). Ward's method minimizes within-cluster variance and thus produces good estimates of cluster groupings. Most of the distance reduction occurred at or before the fifth splitting, and additional splitting resulted in only insignificant changes in variance reduction. Thus, the 24 countries can be adequately described as belonging to only six clusters. Consistent with our rejection of a permissive-nonpermissive dichotomy, we conclude that ISSP countries can be described by six distinctive sexual regimes.

Finally, to determine which counties belong to which of the six clusters, we turned to k-mean algorithms to ascertain the optimal clustering solution for the data (Everitt, 1993). K refers to the number of groups, which we specified to be six (k = 6) on the basis of our hierarchical cluster analysis. We employed PAM, a specialized k-mean program which, unlike other k-mean algorithms, is not sensitive to the ordering of cases in the data set. This program finds the most representative country scores and uses them as starting values or seeds to form the six clusters (Kaufman & Roussecuw, 1990).

Final groupings are presented in Table 2 with information about within- and between-cluster sum of squares. Within-cluster sum of squares indicates the distance between countries in each cluster. The lower the within-cluster sum of squares, the better or more homogeneous is the cluster. Between-cluster sum of squares measures the distance between the cluster profile and the average profile model. A larger between-cluster sum of squares indicates a more distinct cluster, and the higher the sum of the between-cluster sum of squares the more differentiated are the clusters.

Table 2. Clusters, Cluster Country Members, and Between- and Within-Cluster Sum of Squares (SS)

Clusters                       Number of     Between-Cluster SS
                               Countries

Teen Permissives                  5                 .72
 East Germany, West                                (.14)
 Germany, Austria,
 Sweden, Slovenia

Sexual Conservatives              4                 .58
 USA, Ireland,                                     (.15)
 Northern Ireland,
 Poland

Homosexual Permissives            5                 .59
 Netherlands, Norway,                              (.12)
 Czech Republic,
 Canada, Spain

Moderate Residuals                8                 .08
 Australia, Great Britain,                         (.00)
 Hungary, Italy, Bulgaria,
 Russia, New Zealand,
 Israel

Japan                                               .43
                                                   (.43)

Philippines                                         .67
                                                   (.67)

Total                             24               3.07

Clusters                       Within-Cluster SS        Ratio
                                                     (Between S/
                                                     Within SS)

Teen Permissives                     .23                3.09
 East Germany, West                 (.05)
 Germany, Austria,
 Sweden, Slovenia

Sexual Conservatives                 .07                8.07
 USA, Ireland,                      (.02)
 Northern Ireland,
 Poland

Homosexual Permissives               .29                2.02
 Netherlands, Norway,               (.06)
 Czech Republic,
 Canada, Spain

Moderate Residuals                   .48                0.17
 Australia, Great Britain,          (.06)
 Hungary, Italy, Bulgaria,
 Russia, New Zealand,
 Israel

Japan                                --                  --

Philippines                          --                  --

Total                               1.07                2.87

Note. Numbers in parentheses are sums of squares divided by number of countries in each cluster.

Two clusters are singletons, that is, a group which contains only one country. The Philippines constitutes its own cluster because of its extremely conservative attitudes toward all kinds of nonmarital sex. In the Philippines there is a consensus that all forms of nonmarital sex are always wrong. The contrast with other counties is especially striking for premarital sex. Japan is also its own cluster, mostly because of its unusual pattern of wrong only sometimes responses concerning premarital sex. On teenage sex and extramarital sex Japan is much like other nations, although the Japanese are more likely than others to report that homosexual sex is always or almost always wrong (87% compared with 68%, on average, across the 24 countries).

The 22 Western countries cluster in four multi-country groups, which we term the Teen Permissives, Sexual Conservatives, Homosexual Permissives, and Moderate Residuals. Average profiles of the first three clusters are presented in Figure 1.

[Figure 1 ILLUSTRATION OMITTED]

The Teen Permissives' cluster, which encompasses Germany (East and West), Austria, Sweden, and Slovenia, is distinguished by relatively high levels of acceptance of teenage sex and premarital sex. As evidenced by the [between-cluster to within-cluster sum of squares ratio of three, this is a fairly strong cluster. Countries of this cluster are more permissive on teenage sex than others (only 35% of respondents chose always wrong as compared to 58%, on average, across the 24 countries). Those countries are also more accepting of premarital sex. Their scores for extramarital sex and homosexual sex, however, closely match the average profile.

The Sexual Conservatives' cluster, which includes Ireland, Northern Ireland, Poland, and the United States, shows relatively strong disapproval for all types of nonmarital sex. It stands in sharp contrast to the Teen Permissives' regime, which gathered together countries with liberal attitudes toward premarital and teenage sex. For instance, premarital sex is considered always wrong in 28% of the cluster's cases as compared to 17%, on average, for the 24 countries. In fact, the Sexual Conservatives present a U-shaped curve for premarital sex, indicating a polarization of opinion on premarital sex not seen in any other cluster. As for other types of nonmarital sex, Sexual Conservatives chose always wrong at a higher rate than average for all questions: teen sex, 78% versus 58%; extramarital sex, 79% versus 66%; and homosexual sex, 74% versus 59%. That the between-cluster sum of squares is eight times larger than the within-cluster sum of squares demonstrates that this cluster is both homogeneous in its views and quite distinct from the average.

The Homosexual Permissives' cluster consists of the Netherlands, Norway, the Czech Republic, Canada, and Spain. With a sum of squares ratio of two, what distinguishes the countries in this cluster is high levels of acceptance of homosexuality. Countries gathered in this cluster have a profile very similar to the average except for homosexual sex. They accept premarital sex, but reject teen sex and particularly extramarital sex. To a large extent, they present a polarized pattern of response for homosexual sex, which is not to be found in other clusters. This cluster shows the same extent of polarization on homosexuality that the Sexual Conservatives' cluster does on premarital sex.

The Moderate Residuals' cluster contains the remaining eight countries. The low sum of squares ratio for the Moderate Residuals (.017) confirms that they are quite heterogeneous and not very distinguishable from the average profile. In fact, the earlier hierarchical clustering analysis scattered these countries across other clusters, thereby increasing the sum of squares within each of those other three clusters. While this category has no real substantive meaning, it serves to increase the coherence of the other groups, maximizing the between-cluster sum of squares and minimizing the within-cluster sum of squares for the Sexual Conservatives, Teen Permissives, and Homosexual Permissives.

That the Sexual Conservatives, Teen Permissives, and Homosexual Permissives have distinctive regimes of sexual attitudes is apparent in Figure 1. The Teen Permissives are visibly more likely than the other two clusters to see nothing wrong with premarital sex and notably less likely to judge teen sex as always wrong. Sexual Conservatives are more likely to view each type of nonmarital sex as always wrong, while Homosexual Permissives are less likely to condemn homosexual sex.(2)

How Much Sharing?

In addition to knowing the number of regimes for the countries, we have a good picture of the features of those regimes, but we still do not know how different or how similar countries are in terms of their profiles. To find out the degree to which sexual attitudes are shared across countries, we computed the sum of squares across the 24 rows of Table 1, using the "average" response--25%, the result of dividing the 100% total for each variable by four, its number of response categories--as the reference model. We used the deviation of each cell's observed proportion from .25 to compute the total sum of squares for the overall data set. To determine how much is shared by countries (shared sum of squares), we computed the sum of squares between the mean proportion for each response category across the 24 countries and the .25 proportion. Then, we compared the within- and between-clusters sum of squares (Table 2) with those numbers. The within-clusters sum of squares, the between-clusters sum of squares, and the shared sum of squares add up to the total sum of squares.

For the 24 countries, the total sum of squares is 21.7, and the shared sum of squares is 17.56. From Table 2, we know that the between sum of squares and the within sum of squares are 3.07 and 1.07 respectively. Percentages for the different sum of squares are presented in Figure 2.

[Figure 2 ILLUSTRATION OMITTED]

ISSP countries have a good deal in common in terms of their views on nonmarital sex. By our calculations, 81% of the variance in attitudes toward nonmarital sex is shared by the 24 countries while 14% is attributable to the six clusters and 5% to country-specific error. Note that when only 22 countries are considered (excluding Japan and the Philippines), 84% is shared, 10% is regime specific, and 6% is country specific.

DISCUSSION

Drawing on survey data from 24 countries, this paper identifies six distinctive regimes of public opinion regarding the propriety of nonmarital sex. The analysis confirms that sexual attitudes are considerably more complex than a simplistic permissive-nonpermissive typology of countries might suggest: While attitudes toward premarital sex, teen sex, extramarital sex, and homosexual sex are correlated, they relate to one another in complex ways across countries. While this finding is not likely to surprise researchers on sexuality, this paper offers rigorous confirmation for a large set of countries.

Of course, these countries do not constitute a representative sample of countries or cultures, being limited to 24 largely Western and industrialized nations which chose to participate in the 1994 International Social Survey Program. The only two Asian nations, Japan and the Philippines, each form a unique, single-country cluster. This demonstrates that they differ not only from Western countries, but also from one another. The Philippines is set apart by particularly conservative attitudes toward all forms of nonmarital sex. For Japan, it is the respondents' propensity to label premarital sex as only sometimes wrong rather than expressing strong approval or disapproval that makes it a singleton. We are hard put to interpret this unique pattern, but it is consistent with one scholar's observation that the Japanese view premarital sex with concern only when it is indiscreet or conflicts with social responsibilities (White, 1993). It might also reflect the fact that the Japanese avoid extreme response categories (Hui & Triandis, 1989), although the other three types of nonmarital sex do not elicit such ambivalent responses.

As for the Western countries, cluster analysis identifies four sexual regimes: Sexual Conservatives, Teen Permissives, Homosexual Permissives, and Moderate Residuals. Since the cluster analysis shows the United States to belong to the Sexual Conservatives, this study confirms that American sexual values are strikingly nonpermissive. The conservative nature of American moral views about sexuality has long been recognized. Pointing to the Europeans' tolerance of premarital relations and to high rates of extramarital sex reported by European-born informants in the United States, Kinsey, Pomeroy, and Martin (1948) cautioned against pooling cross-national data for the United States and Continental Europe. A half century later, ISSP survey evidence shows Kinsey was correct. The United States is distinctive--markedly more conservative than most European nations in its sexual attitudes. In fact, American attitudes toward nonmarital sex class it with Ireland, Northern Ireland, and Poland--three nations associated with conservative Catholic populations. Other nominally Catholic countries (e.g., Italy) do not fall into this cluster, so the common denominator of the Sexual Conservatives is not religion per se. An analysis of ISSP data on church attendance, however, does show that the countries grouped as Sexual Conservatives are those where people report attending church frequently.

The cluster analysis also shows that there is no monolithic Continental attitude toward sexuality in Europe, but rather diversity and nuances in European moral beliefs about premarital sex, teen sex, extramarital sex, and homosexual sex. Besides the Sexual Conservatives, there are two other strong clusters of Western countries: the Teen Permissives and the Homosexual Permissives, in addition to the Moderate Residuals, who fall somewhere in between other clusters. One cluster of countries is relatively tolerant of premarital sex and even teen sex. Another cluster is distinguished by comparatively high acceptance of homosexuality.

As for the Teen Permissives' cluster, researchers have viewed Sweden as unusually accepting of sex before marriage (Linner, 1967) for decades. Even acceptance of teenage sexuality is regarded as "unremarkable in a country like Sweden, with its long history of support for sexual freedom, and the absence there of taboos against premarital sex" (Jones et al., 1985, p. 59). The cluster analysis shows that Sweden is not uniquely permissive in terms of approval of premarital sex. As indicated by their membership in the Teen Permissives' cluster, Germany (East and West), Austria, and Slovenia hold views about premarital sex and teenage sex which are very similar to those of Sweden. If illegitimate fertility--high by European standards--may be taken as an indication of tolerance of sex before marriage, then Sweden, Germany, and Austria were all relatively permissive regarding premarital sex even at the end of the nineteenth century (Knodel, 1974).

The Netherlands, Norway, the Czech Republic, Canada, and Spain are distinguished by their tolerance for homosexual sex, but there are differences within the Homosexual Permissives' cluster as well. Most countries which register high levels of tolerance toward homosexuality are also deeply divided. While big shares of the population see nothing wrong with same-sex relations, significant segments consider homosexuality to be always wrong. The exception is the Netherlands, where the majority of residents have come to accept the more tolerant views. A number of factors have been cited to account for Dutch open-mindedness on sexual matters. Although opposition of conservative religious elements deters broad acceptance of homosexuality in many countries, Oosterhuis (1996) notes that Dutch Christian communities have been a progressive force in redefining homosexuality from a sin to a personal issue warranting Christian compassion. To explain permissive Dutch attitudes, others have commented on the unusually long history of the Dutch homosexual movement (Tielman, 1987), the self-conscious pluralism of the Netherlands, the secularism of its population, and the candor and coverage of sexual topics in the mass media (Jones et al., 1986).

While it is valuable to know which countries cluster together, it is also revealing to note which countries adopt different sexual regimes. As we indicated, there is no distinctive regime that characterizes Catholic countries. Nor does a history of socialism lead to a singular view of nonmarital sexuality: In fact, formerly communist countries are to be found in each of the four clusters of Western nations. If there is an Anglo-Saxon tradition of sexual attitudes, it extends to Great Britain, Australia, and New Zealand, but not to the United States, Canada, or Ireland. Although Scandinavian countries are generally assumed to share a similarly tolerant attitude toward sexuality (Jones et al., 1986), Norway is seen to be less tolerant of premarital sex, especially young teen sex, and more approving of homosexuality than is Sweden. Indeed, the results of our study serve to place findings for individual countries in a broader comparative context, which helps to evaluate popular beliefs about particular countries and to point out new research issues. For example, on the basis of a careful analysis of the best data available to him, the distinguished Russian sex researcher Igor Kon (1995) concluded that Russian sexual values were "very similar to those of the Western world" (p. 177). While our analysis generally confirms this for attitudes toward premarital sex, teen sex, and homosexual sex, Russians' greater tolerance of extramarital sex warrants further consideration.

Although this study reveals different normative regimes, it also points out that there is much sharing going on between countries. Compared to cognitive maps of emotions and kinship terms (Romney, Boyd, Moore, Batchelder, & Brazill, 1996; Romney, Moore, & Rusch, 1997), the 81% of variance in sexual attitudes which is shared can be regarded as high. This degree of sharing reflects the fact that cultures around the globe have been shown to make use of only a small proportion of the logically possible combination of sexual norms (Heise, 1967). Because our analysis focuses on a relatively homogeneous subset of largely industrialized and Western nations in an increasingly globalized world, the potential for overlap in moral judgments about sex is undoubtedly high. In general, these 24 nations take a cautious stance toward all types of nonmarital sex except premarital relations among adults.

Despite considerable agreement across cultures on the rightness and wrongness of nonmarital sex, the cluster analysis finds six distinct moral regimes typifying the 24 countries examined. Leaving aside the distinctive profiles of Japan and the Philippines, Western nations can he grouped with other countries that share their general conservatism, their acceptance of premarital sex, or their tolerance toward homosexuality. These sexual regimes exist even though pluralistic and contradictory values about sex may flourish within national populations, as seen most clearly in the polarized attitudes toward homosexual sex and premarital sex in several nations. Across nations, extramarital sex (which is generally disapproved) generates the most agreement, while attitudes toward homosexual sex and teen sex differentiate countries and clusters.

(1) The data utilized in this paper were documented and made available by the Zentralarchiv Fuer Empirische Sozialforschung, Koeln. The data were collected for the ISSP by independent institutions in each country. Neither the original collectors nor the Zentralarchiv bear any responsibility for the analyses or interpretations presented here.

(2) Let us underline that the clusters do not correlate with questionnaire types. Comparing results of Table 2 with Appendix A, one can see that modes of data collection are evenly distributed across clusters.

REFERENCES

Bozon, M., & Leridon, H. (1996). Sexuality and the social sciences: A French survey on sexual behavior. Aldershot, UK: Dartmouth.

Broude, G., & Greene, S. J. (1976). Cross-cultural codes on twenty sexual attitudes and practices. Ethnology, 15, 409-429.

Brown, J. S. (1952). A comparative study of deviations from sexual mores. American Sociological Review, 17, 135-146.

Buss, D. M. (1989). Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behavioral and Brain Sciences, 12, 1-14.

Carballo, M., Cleland, J., Carael, M., & Albrecht, G. (1989). A cross national study of patterns of sexual behavior. The Journal of Sex Research, 26, 287-299.

Christensen, H. T., & Carpenter, G. R. (1962). Timing patterns in the development of sexual intimacy: An attitudinal report on three modern Western societies. Marriage and Family Living, 24, 30-35.

Christensen, H. T., & Gregg, C. F. (1970). Changing sex norms in America and Scandinavia. Journal of Marriage and the Family, 27, 616-627.

Cleland, J., & Ferry, B. (1995). Sexual behaviour and AIDS in the developing world. London: Taylor & Francis.

Everitt, B. S. (1993). Cluster analysis. New York: Edward Arnold.

Ford, C. S., & Beach, F. (1951). Patterns of sexual behavior. New York: Harper and Row.

Heise, D. R. (1967). Cultural patterning of sexual socialization. American Sociological Review, 32, 726-739.

Hui, C., & Triandis, H. (1989). Effects of culture and response format on extreme response style. Journal of Cross-Cultural Psychology, 20, 296-309.

Inglehart, R. (1990). Culture shift in advanced industrial society. Princeton, NJ: Princeton University Press.

Inglehart, R. (1997). Modernization and postmodernization. Princeton, NJ: Princeton University Press.

Jones, E. F., Forrest, J. D., Goldman, N., Henshaw, S. K., Lincoln, R., Rosoff, J. I., Westoff, C. F., & Wulf, D. (1985). Teenage pregnancy in developed countries: Determinants and policy implications. Family Planning Perspectives, 17, 53-63.

Jones, E. F., Forrest, J. D., Goldman, N., Henshaw, S., Lincoln, R., Rosoff, J. I., Westoff, C. F., & Wulf, D. (1986). Teenage pregnancy in industrialized countries. New Haven, CT: Yale University Press.

Kaufman, L., & Rousseeuw, P. J. (1990). Finding groups in data: An introduction to cluster analysis. New York: Wiley-Interscience Publications.

Kinsey, A. C., Pomeroy, W. B., & Martin, C. E. (1948). Sexual behavior in the human male. Philadelphia: W.B. Saunders Co.

Knodel, J. E. (1974). The decline of fertility in Germany, 1871-1939. Princeton, NJ: Princeton University Press.

Kon, I. (1995). The sexual revolution in Russia. New York: The Free Press.

Kyes, K. B., & Tumbelaka, L. (1994). Comparison of Indonesian and American college students' attitudes toward homosexuality. Psychological Reports, 74, 227-237.

Laumann, E. O., Gagnon, J. H., Michaels, R. T., & Michaels, S. (1994). The social organization of sexuality: Sexual practices in the United States. Chicago: University of Chicago Press.

Levi-Strauss, C. (1969). Les regles elementaires de la parente [The elementary structures of kinship]. Boston, Beacon Press.

Linner, B. (1967). Sex and society in Sweden. New York: Pantheon House.

Lottes, I. L. (1985). The use of cluster analysis to determine belief patterns of sexual attitudes. The Journal of Sex Research, 21, 405 421.

Lottes, I. L., & Weinberg, M. S. (1997). Sexual coercion among university students: A comparison of United States and Sweden. The Journal of Sex Research, 34, 67-96.

Ma, L. (1989). Premarital sexual permissiveness of American and Chinese college students. Sociological Spectrum, 9, 285-299.

Minturn, L., Grosse, M., & Haider, S. (1969). Cultural patterning of sexual beliefs and behavior. The Journal of Sex Research, 8, 310-318.

Murdock, G. P. (1960). Social Structure. New York: Macmillan and Co.

Oosterhuis, H. (1996). Christian social policy and homosexuality in The Netherlands, 1900-1970.Journal of Homosexuality, 32, 95-111.

Perlman, D., Josephson, W., Hwang, W. T., Begum, H., & Thomas, T. L. (1978). Cross-cultural analysis of student's sexual standards. Archives of Sexual Behavior; 7, 545-558.

Romney, A. K., Boyd, J., Moore, C., Batchelder, W., & Brazill, T. (1996). Culture as shared cognitive representation. Proceedings of the National Academy of Sciences, 93, 4699-4705.

Romney, A. K., Moore, C., & Rusch, C. (1997). Cultural universals: Measuring the semantic structure of emotion terms in English and Japanese. Proceedings of the National Academy of Sciences, 94, 5489-5494.

Ross, M. W. (1985). Actual and anticipated societal reaction to homosexuality and adjustment in two societies. The Journal of Sex Research, 21, 40-55.

Schwartz, I. M. (1993). Affective reactions of American and Swedish women to their first premarital coitus: A cross-cultural comparison. The Journal of Sex Research, 30, 18-26.

Smith, T. W. (1992a). Discrepancies between men and women in reporting number of sexual partners: A summary from four countries. Social Biology, 39, 205-211.

Smith, T. W. (1992b). The International Social Survey Program. International Journal of Public Opinion Research, 4, 275-278.

Sprecher, S., & Hatfield, E. (1996). Premarital sexual standards among U.S. college students: Comparison with Russian and Japanese college students. Archives of Sexual Behavior, 25, 261-288.

Sprecher, S., Hatfield, E., Cortese, A., Potapava, E., & Levistskaya, A. (1994). Token resistance to sexual intercourse and consent to unwanted sexual intercourse--College students' dating experiences in 3 countries. The Journal of Sex Research, 31, 125-132.

Tielman, R. (1987). Dutch gay emancipation history (1911-1986). In: A. X. Van Naerssen (Ed.), Gay life in Dutch society (pp. 9-17). New York: Hurrington Park Press.

Weinberg, M. S., Lottes, I. L., & Shaner, F. M. (1995). Swedish or American heterosexual youth: Who is more permissive? Archives of Sexual Behavior, 24, 409-438.

Wellings, K., Field, J., Johnson, A. M., & Wadsworth, J. (1994). Sexual behavior in Britain: The national survey of sexual attitudes and lifestyles. London: Penguin Books.

White, M. I. (1993). The material child: Coming of age in Japan and America. New York: Free Press.

Manuscript accepted March 18, 1998

Appendix A. Description of the national samples

Countries          Sample Size   % Female   Mean Age

Australia          1779          49         49.5
Austria             977          56         46.3
Bulgaria           1126          60         48.3
Canada             1440          63         38.1
Czech Republic     1024          55         41.5
Germany (East)     1097          52         47.0
Germany (West)     2324          49         45.2
Great-Britain       993          55         46.8
Hungary            1500          58         48.7
Ireland             938          52         45.7
Israel             1287          55         41.1
Italy              1021          51         42.7
Japan              1306          55         41.3
Netherlands        1968          55         43.5
New Zealand        1047          59         45.8
Northern Ireland    647          53         46.6
Norway             2087          55         40.4
Philippines        1200          50         38.3
Poland             1597          55         45.6
Russia             1998          64         40.8
Slovenia           1032          54         42.8
Spain              2494          52         44.4
Sweden             1272          53         44.8
USA                1447          59         45.9

Countries          Questionnaire Mode

Australia          Mail
Austria            Face to face
Bulgaria           Face to face
Canada             Self-completion
Czech Republic     Face to face
Germany (East)     Self-completion
Germany (West)     Self-completion
Great-Britain      Self-completion
Hungary            Face to face
Ireland            Face to face
Israel             Face to face
Italy              Face to face
Japan              Face to lace
Netherlands        Face to face
New Zealand        Mail
Northern Ireland   Self-completion
Norway             Mail
Philippines        Face to face
Poland             Mail
Russia             Self-completion
Slovenia           Face to face
Spain              Face to face
Sweden             Mail
USA                Self-completion

Eric D. Widmer, Judith Treas, and Robert Newcomb

University of California-Irvine

Address correspondence to Eric D. Widmer, Department of Anthropology, University of California-Irvine, Irvine, CA 92697-5100; e-mail: [e-mail]