Chapter 3

Racial and Sexual Differences

Introduction to the Analysis of Tables

The response to the statements in the male and female scales noted in Chapter 2 makes it quite clear that Negroes and males are respectively more permissive than whites and females. This relationship holds up in both the student and adult samples. This chapter will explore some correlates of these differences.1
One of the significant findings of this study is that racial and sexual differences are not simply differences in degree but reflect a rather basic difference in the sociocultural situation of these groups. Other studies have pointed out the greater permissiveness of Negroes.2 However, most of these are intensive psychological studies or participant-observation studies of a community. The few that are more like surveys report little more than that this greater permissiveness exist. Race and sex differences were elaborated, and some data on this from the present study were presented, in several of the tables in Chapter 2. In particular, the relation of these race and sex groups to certain religious and courtship institutional structures is examined in this chapter. This should begin to explain the crux of the differences between these race and sex groups.
The tests reported on in this chapter (and in most of the other chapters) use the six-scale-type universally ordered scale, and as noted in Chapter 2 (see Table 2.14), the respondents are characterized by their responses to the scale that corresponds to their own sex. The tables show this scale dichotomized between scale-types two and three, or divided into respondents who accept coitus under some condition and those who do not accept coitus under any condition. This is approximately the place that is suggested by intensity analysis as a suitable cutting point. (See Appendix D.) It should be clear that the test for almost all the tables computed in this and all other chapters were run with different and more-varied cuts in order to be sure of the linear quality of the relationship and in order to be sure that the results were not due simply to an arbitrary cut of the scale.3
As mentioned earlier, the method of analysis employed was that of partial tables. Simply put, this means that "controls" such as race were used and that the resulting partial tables were compared to see if the partial table for whites differed from that for Negroes. In this way the relation of various factors to premarital sexual permissiveness could be examined, and a check could be made to see if the relation held up using such a control as race.
The usual way to decide if two partial tables differ from each other is to see if the relationship being examined is significant for one category of the control but not for the other. For example, it might be found that upperclass people are significantly less permissive than lower-class people when looking at only whites but that when checking Negroes the class difference is not significant with regard to permissiveness. In such a case the conclusion would be that social class relates to permissiveness only for whites. A problem with this approach is that it is possible for one partial table to be just above the level of significance and for another to be just below the level of significance. In such a case, the actual difference is slight, and yet the conclusion would be that the relationship held up in one control category but not in the other. A related second problem revolves around the fact that the number of cases affects the size of the chi-square, so that if there were many more whites than Negroes, the chi-square would tend to be larger for whites. The use of gamma as a measure of the strength of the relationship helps to resolve this second problem in these cases, for it affords a better comparison of tables with unequal numbers of cases. However, the basic problem remains, for it is necessary to have a method of deciding how large a difference in gammas is necessary before it can be said that the two partial tables are significantly different.
In 1964, Leo Goodman published an article that presented a method of solving this problem in the analysis of partial tables.4 This method yields a figure that enables one to conclude whether the difference between two partial tables is large enough to be significant at the .05 level or at any other level. Simply put, this method affords a way of doing more than just checking the chi-square or gamma in two tables. It affords a more precise way of discerning how likely it is that the difference between two tables is due to chance.
One illustration of the value of this new test for interaction can be seen by looking at Table 3.4. By comparing the two tables for high church-attending females it appears that one is significant at the .05 level and that the other is not. However, a comparison of the gammas shows no difference, and a check using the Goodman technique indicates that the relationship of romantic love to permissiveness is not significantly different for white and Negro females. Thus, the conclusion is something quite different than what a rote use of the chi-square of each partial table would have indicated. This method adds another element of rigor, another check, and thereby affords a sounder base for the interpretation of partial tables. Thus it is used throughout this study as an additional check.

Preface to Proposition One

In order to clarify the presentation, the proposition summarizing the findings reported in each of the next seven substantive chapters is informally presented early in each chapter. However, the presentation will make it amply clear that in reality the proposition came last, after the empirical findings were digested.
Basically, the examination of race and sex differences discussed in this chapter indicated that in their degree of permissiveness white females showed a considerable amount of sensitivity to social forces such as religiosity, romantic-love beliefs, and the number of times they had been in love. Negro males showed very little relation between their permissiveness and any of these social forces. Generally speaking, male permissiveness was affected less by such social forces than was female permissiveness, and Negro permissiveness was affected less than was white permissiveness.
These empirical findings led to the proposition that it was the permissiveness of groups with a tradition of low sexual permissiveness (whites and females) that was more capable of being altered by social forces than was the permissiveness of groups with a tradition of relatively high sexual permissiveness (Negroes and males). In other words, it seems that the likelihood of a change in sexual permissiveness is greater for groups who are low on permissiveness, and this implies that the change would be toward more permissiveness, since these low permissive groups are already quite low on permissiveness. The findings, in detail, and a more formal statement and analysis of the proposition follow.

Race and Permissiveness

The relationship of race to permissiveness by sex is presented in Table 3.1, and it is clearly strong and significant. Although the general level of permissiveness for all groups is lower in the adult sample, race and sex differences are present in the adult sample in slightly stronger form than in the student sample.
The difference in level of permissiveness by sex and race is important, but there are other associated differences that support the position that this is more than just a difference in degree of permissiveness. The relation of several social factors to the level of permissiveness in these sex and race groups differs significantly. The tests concerning these factors were carried out only in the student sample.

Table 3.1
RACE AND PERMISSIVENESS BY SEX
IN THE STUDENT AND ADULT SAMPLES
(PERCENTAGE HIGHLY PERMISSIVE)a

        MALE            FEMALE
Student Sample
White   61 (287)*       27 (324)
Negro   85 (115)        45 (118)
        c2 = 21.6       c2 = 13.6
        P <.001              P <.001
        Q=.57**         Q=.39
                
Adult Sample
White   30 (607)        6 (649)
Negro   65 (62)         30 (81)
        c2 =30.2        c2 =49.6
        P <.001              P <.001
        Q=.62           Q=.73

a The permissiveness of each sex is measured by rcsponses to the scale of the same sex; that is, male permissiveness by the answers to the male scale, and female permissiveness by the answers to the female scale.
* In this and all following tables the number in parentheses is the base for the percentage.
** Q is the same as gamma in 2-by-2 tables and is a measure of the strength of the relationship.

Religious Attendance and Permissiveness

Kinsey and others have demonstrated the connection between religious behavior and premarital sexual permissiveness.5 Accordingly, it was hypothesized for this study that the more religious elements would be lower on permissiveness. No racial differences were hypothesized. The hypothesis was tested only in the student sample, using several measures of religiosity. However, for the purpose of this discussion church-attendance is used as the measure of religiosity, although other measures showed very similar results (Appendix A, Part I, question 5c). High church-attendance was defined as going to church more than once a month. Different cutting points showed the same result. It is worth noting that this result, as well as all results on the student sample, were checked in each of the five schools. Thus, there are really five independent samples that test the findings given for the total student sample.6
The zero-order relation between church-attendance and permissiveness negative and significant.7 Table 3.2 shows that when this relation is

Table 3.2
CHURCH ATTENDANCE AND PERMISSIVENESS BY RACE AND SEX
IN THE STUDENT SAMPLE (PERCENTAGE HIGHLY PERMISSIVE)

CHURCH-               Male               Female
ATTENDANCE      WHITE     NEGRO     WHITE     NEGRO
Low             77 (159)  91 (33)   53 (136)  58 (12)
High            40 (121)  83 (81)   6 (183)   44 (105)
                c2 =41.0  c2 =1,2   c2 =92.1  c2 =0,9
                P <.001   NS     P <.001   NS
                Q=-.68    Q=-.35    Q=-.90    Q=-.29

controlled by race and sex it becomes evident that the relation holds more for whites than for Negroes, and that within the white group it holds more for women than for men. The differences just noted appear valid, for they are sizable, are supported in all schools and are not altered by the use of different check variables. All three white schools in the five-school sample showed the negative relation of permissiveness to church-attendance, while neither of the two Negro schools did so. Since the white New York college is highly permissive, it seems that the racial difference found is not simply due to Negroes being more permissive but most likely reflects some genuine differences in the Negro subculture.
The stronger association for white females is congruent with other researchers' findings that religion exerts more control over the female's sexual life than it does over the male's. The culturally less-developed erotic imagery of females may make control of their sexual life easier.
Goodman's test for Interaction generally supports this interpretation of Table 3.2. For example, the differences between white and Negro females in the relationship of church-attendance to permissiveness are significant at the .01 level. However, the differences between white and Negro males are significant only at the .25 level, indicating a weaker male difference. The difference in this relationship between white males and white females is significant at the .01 level, 8 but the difference for Negro males and females is not significant.
The Negro-white difference in the association of church-attendance and permissiveness can be seen another way by shuffling the columns in Table 3.2 If the relation between permissiveness and race is controlled for by

Table 3.3
RACE AND PERMISSIVENESS BY SEX AND CHURCH ATTENDANCE IN THE STUDENT
SAMPLE (PERCENTAGE HIGHLY PERMISSIVE)

        High Church-Attendance          Low Church-Attendance
RACE    MALE      FEMALE                MALE      FEMALE
White   40 (121)  6 (183)               77 (159)  53 (136)
Negro   83 (81)   44 (105)              91 (33)   58 (12)
        c2 =36.7  c2 =62.7              c2 =3.1   c2 =0.1
        P <.001   P <.001               P <.10   NS
        Q = .76   Q = .86               Q = .49   Q = .11

church-attendance, then the racial difference is reduced. In this table the Negro-white difference in permissiveness is accentuated among the high church-attenders but reduced considerably among those who were the low church-attenders. This result is due to the fact that the negative association between church-attendance and permissiveness holds predominantly for whites. Thus, low church-attending whites show a significant increase in permissiveness, while the low church-attending Negroes show relatively little increase in permissiveness; as a result, the racial difference tends to disappear in this group of low church-attenders. Among the high church attenders, the racial difference is accentuated and strong racial differences become obvious. This result further indicates that church-attendance implies something rather different in Negro culture than it does in white culture.
In the latter it seems to be symbolic of a conservative style of life. It is a key variable in understanding premarital sexual permissiveness and is discussed frequently in this study.

Romantic Love and Permissiveness

Beside the investigation of racial difference in level of permissiveness and the way religion affects permissiveness, an investigation was made of the relation of romantic-love beliefs to permissiveness, using the student sample. An eight-item Guttman scale was employed to measure romantic love.9 (This is discussed and analyzed in more detail in Chapter 5.) For purposes of this analysis the respondents were divided into the categories high and low, on the basis of whether or not they endorsed item 4: There is only one real love for a person. The results were quite similar even if the entire scale was used with more cuts or with different cuts.
Just as with the previous examination, no racial differences were expected, but they did occur. It was assumed that both races would show that those who were low on romantic love were more permissive sexually. It was reasoned that romantic love was an idealistic, conservative element in our culture, and that the less an individual adhered to this belief the more permissive he would be. Nevertheless, the zero-order relation between romantic love and permissiveness was weak and not significant. Among white students, however, the relationship was significant and negative, as was hypothesized. On the other hand, in the Negro group the association was positive, though slightly short of the .05 level of significance. Both the negative association among white students and the positive association among Negro students were stronger for women than for men.
Since church-attendance had shown itself to be a powerful influence, under certain conditions, it was used here as a control on the relationship of sexual permissiveness to romantic love. It was found that the positive association between romantic love and permissiveness among Negro women held largely for those who were high on church-attendance, whereas the negative association among white women held primarily for those low on church-attendance. Males generally showed no effect of this control and continued to exhibit relations that were not significant. However, high church-attending Negro males did show some tendency toward a positive relationship.
Church-attendance proved to be the most powerful control, more powerful than race and much more powerful than sex. This can be seen in Table 3.4. High church-attendance interacts with the relationship between

Table 3.4
ROMANTIC LOVE AND PERMISSIVENESS BY RACE, SEX, AND CHURCH-ATTENDANCE
IN THE STUDENT SAMPLE (PERCENTAGE HIGHLY PERMISSIVE) [a]

                High Church Attendance  Low Church Attendance
ROMANTIC LOVE   WHITE    NEGRO          WHITE    NEGRO
        Female
High            9 (57)   58 (45)        30 (27)  50 (2)
Low             3 (124)  33 (55)        58 (106) 60 (10)
                c2 =2.6  c2 =6.3        c2 =6.7 
                NS       P <.05      P <.01 P = .85*
                Q=.48    Q=.48          Q=-.53   Q=-.20

        Male
High            42 (36)  93 (27)        70 (30)  90 (10)
Low             40 (78)  79 (47)        80 (123) 94 (18)
                c2 = 0.0 c2 = 2.4       c2 = 1.3        
                NS       NS             NS       P = .59*
                Q=.04    Q=.54          Q=-.25   Q=-.31

a. Correction for continuity could be used in the smaller not-significant tables. This would of course merely lower the already not-significant chi-square. It is used in other fourfold tables where the correction could make a difference in significance.
* Fisher's "Exact" test was used instead of chi-square because of the small number of low church-attending Negroes.

romantic love and permissiveness in all sex and race groups to produce some sort of positive association of these variables, even though this is significant only for Negro females. Again, low church-attendance interacts with the relationship between romantic love and permissiveness to produce some sort of negative association of these variables, even though this is significant only for white females. It is difficult to throw away all the nonsignificant groups arbitrarily, for the difference between them and the two significant groups is not always great. Instead, it would seem best to conclude that high church-attenders generally tend to display a positive association between romantic love and permissiveness, and low church-attenders generally tend to display a negative association, and that this is generally stronger for females than for males. This is a more reasonable statement than simply to assert that the romantic love-permissiveness relation holds exclusively for certain racial groups of females. Finally, it might be noted that here again a variable-romantic love-has greater effect on females than on males. This might be due to the absence of strong sexual motives among females, which aids variables like romantic love in affecting sexual permissiveness.
Goodman's test for interaction generally supports this interpretation. It is interesting to note that although white and Negro females in both high and low church-attending groups display an interaction effect (with only Negro females significant in the high church-attendance group and only white females significant in the low church-attendance group) the test for interaction shows no significant difference in the relation of romantic love to permissiveness between white and Negro females in the high church-attendance group and no significant difference between the white and Negro females in the low church-attendance group. This agrees with the previously stated interpretation that all females show the same tendencies when in the same church-attendance categories. The low church-attenders seem to define romantic love differently than do the high church-attenders. Most Negroes happen to be high church-attenders and thus display the positive association, but the tendency also is there among white females. One value of this check for interaction is that it affords the opportunity to test for a significant difference between partial tables and thus avoids mere mechanical use of them. 10

Frequency of Falling in Love and Permissiveness

One additional test was made-it was on the hypothesis that those who have been in love more often are more permissive. This was expected to be particularly true for females, for research has shown that a woman's sexual life is more affected by being in love than is a man's.11 This was checked in the student sample. The zero-order relationship was significant but weak (c2 = 11.6 and gamma = .20), and it held somewhat more strongly among women than men. The association was strongest among white women and just reached the level of significance among white men, whereas in both Negro sexes it was weak and not significant. Once again the data showed a racial difference, and once again such a state of affairs had not been hypothesized.
The relationship among white students of permissiveness to the number of times an individual has been in love differs according to sex, for as the number of times a woman has been in love increases so does her permissiveness; whereas among men the relationship is curvilinear, with those who have been in love only once being less likely to be permissive than those who have never been in love or who have been in love twice or more. A control by church-attendance does not alter these relationships, but romantic love (as can be seen in Table 3.5) does prove to be a condition of the relationship. Women low on romantic love and men high on romantic love showed a positive relation between the number of times they had been in love and permissiveness. Although the relation was not significant, white men who were low on romantic love showed the curvilinear relation that characterized the entire group of white

Table 3.5

males. White women showed a positive relation regardless of the degree of romantic love, although the relation was stronger for women low on romantic love. Although the Negro groups did not show any significant relation, it was clear from the results of their tests that they showed some tendencies in the same direction as the whites. However, tests using different measures and different cuts in each school substantiated the racial difference.
The Goodman test for interaction as applied to the partial tables in Table 3.5 generally supports the interpretation just given, but it does qualify the strength of the differences noted. Romantic love as a control seems to affect males somewhat more than females in that white males who differ on romantic love are significantly different; so are Negro males. However, females do not show significant differences. This is so despite the fact that on low romantic love white females show a significant relation and Negro females do not. The difference here is not significant, and this demonstrates the point that two partial tables may not be significantly different even though one is significant and the other is not. Finally, the male-female differences in this table generally seem to be stronger than the Negro-white differences.

Summary, Conclusions, and Proposition One

The Negro-white differences found in all the examinations detailed in this chapter were largely unexpected. Of course, they can be explained in a very general sense as the result of different historical backgrounds, which have produced contrasting orientations toward the church, romantic love, falling in love, and sexual permissiveness in general. This explanation is supplemented here with a less abstract and more specific factor-namely, the traditional level of premarital sexual permissiveness in a group.

The relationships discussed were checked carefully in order to be sure they were not due to the arbitrary use of a type of cutting point or to a particular way of measuring social forces. Results were also checked whenever possible to ascertain if they held up in each of the five schools. The highly permissive white New York college was compared with the Negro schools to see if equally permissive schools would still show a racial difference. They did. The Negro-white differences were not due to social class, for in tests in which class was controlled for, the differences remained quite strong. Twelve different measures of social class were used to check this result.12 The Negro-white differences in courtship reported here are but part of a larger set of differences (presented in other chapters) and thus are not a rare finding.

Table 3.6

By reading across Table 3.6, which presents the race comparison for both the student and adult samples, one ean see that with social class controlled for, Negro-white differences in permissiveness persist in both samples. Dollar income and Duncan's Socioeconomic Index (SEI) are used to measure social class. There are few upper-class Negroes, and thus, the best comparisons with whites can be made at the lower levels. It is clear that at the lower levels both male and female Negroes are more permissive than whites. Other checks using different cuts show the same results. All the relations discussed in this chapter remain unaltered when they are computed with a control for social class. Despite the popularity of assuming that social class is the great equalizer of all race differences, this is not the case in these data.13
Negro sexual permissiveness seems to have different sources and different implications. For example, students in the highly permissive white New York college were as permissive as those in the Negro schools, but the majority of these white students were low on church-attendance and romantic-love beliefs, whereas the Negro students generally were high on these variables. It seems that a liberal, or permissive, attitude toward premarital sexual behavior is generated and maintained in a different manner in Negro subcultures than in white subcultures.
In looking over these Negro-white differences and searching for an explanation, the similarity of the differences between men and women within each racial group and between Negroes and whites considered as total groups was striking. To illustrate-men are significantly more permissive than women; romantic love does not affect men in the same way it does women; and among white students church-attendance and the number of times the individual has been in love affects female permissiveness more than male permissiveness. Very similar differences were reported between Negroes and whites. This situation reflects a general difference between subcultures in American society that are traditionally more, and traditionally less, permissive. The finding that the social forces examined in this study affected white permissiveness more than Negro permissiveness and female permissiveness more than male permissiveness suggests that an inverse relation exists between the traditional level of permissiveness and the susceptibility of an individual's permissiveness to such social forces.
Proposition One subsumes these sex and race differences as special cases: THE LOWER THE TRADITIONAL LEVEL OF SEXUAL PERMISSIVENESS IN A GROUP, THE GREATER THE LIKELIHOOD THAT SOCIAL FORCES WILL ALTER INDIVIDUAL LEVELS OF SEXUAL PERMISSIVENESS.
The findings fit this theoretical explanation rather well. Studies of the Negro church indicate that it is a source of emotional satisfaction rather than an inhibitory influence on sexual behavior.14 The church may in fact strive to reduce sexual permissiveness, but the strong tradition of sexual permissiveness among Negroes seems to counteract this effort, whereas the less-permissive white customs do not have this counteracting force. Similarly, church-attendance seems less able to affect the sexual attitudes of white males as compared to white females; the stronger tradition of sexual permissiveness in males seems more able to counteract the religious efforts at control.
Church attendance also showed its greater effect on females in the relationship of romantic love to permissiveness. High church-attenders showed a positive relation, and low church-attenders showed a negative relation, of romantic love to permissiveness predominantly among females. The males showed no significant relation. The Negro-white differences here were minor and mainly in the female group, where the greater church-attendance adds an emotional quality to the idealistic aspects of romantic love that makes those who are high on both more likely to accept coitus. Perhaps the weaker response of males was due here to the fact that their traditionally high acceptance of permissiveness was sufficient motivation to sexual acceptance, and thus they had no need for such "stimulants" as romantic love.
The relation between permissiveness and the number of times an individual has been in love is affected by the same racial and sexual differences. This relationship held only for white students. It can be argued here too that as a consequence of their traditional acceptance of sexual permissiveness Negroes do not "need" to fall in love as much as whites do in order to promote permissive sexual behavior.15 The same argument holds for the difference between males and females in the white group, and it explains why the men among whom the relationship between permissiveness and the number of times they have been in love is strongest are a small group of men high on romantic love and somewhat low in sexual permissiveness.
Thus, within traditionally less-permissive groups-women and whites-individual permissiveness is more likely to be affected by such social forces as church-attendance, belief in romantic love, and falling in love.16 In the traditionally more-permissive groups-men and Negroes-individuals find support and justification for liberal sexual attitudes in the groups' traditions, and their permissiveness is therefore less subject to control by social forces. Highly permissive individuals in traditionally low-permissive social groups are permissive not because they have long-standing traditions to support them, but because they are located in the social structure in such a way as to avoid inhibitory forces (church attendance and the idealistic version of romantic love for example) and to maximize experiences that promote permissiveness (like falling in love). The best illustration of this is the highly permissive white college students. They were consistently low on church-attendance, tended not to believe in romantic love, and reported falling in love relatively often. The best example of a traditionally high-permissive group is the Negro students. They lacked these characteristics, their permissiveness seeming rather to be a consequence of a long-standing supportive tradition. Thus, there are two distinct paths, both of which lead to high individual levels of permissiveness. The key differentiating factor is the traditional level of premarital sexual permissiveness in the group to which the individuals belong. In this sense, although the white New York college and the Negro college are similar in levels of permissiveness, they are quite distinct in the ways in which these levels were achieved and maintained.
This characterization of the difference between traditionally more- and less-permissive subcultures is further supported by the contrast between two extreme groups-Negro men and white women. The permissiveness of white women is affected by all the variables investigated; that of Negro men, by none of them.
The theory of social change in sexual permissiveness points to several leads for future research and theory. To evaluate the efficacy of any variable on sexual permissiveness it is vital to know the traditional level of permissiveness of the group. A variable may seem relatively impotent in reducing permissiveness in a traditionally high-permissive group, whereas in a traditionally low-permissive group it may have great effect. The theory also suggests that differences in the area of sexual permissiveness between races are analogous to the differences between the sexes-studying one set may well be essential to understanding the other. Finally, Proposition One implies a strong tendency toward long-range, unidirectional change. Once a group becomes highly permissive and stays that way long enough for it to become traditionally so, then it becomes quite difficult for social forces to reduce that permissiveness. On the other hand a group with a tradition of low permissiveness can be altered either way. Since such a group is already low on permissiveness, there is little room to move down, and there is a good chance of increasing the traditional level of permissiveness. The implications are that often it should be possible to find in history a longrange trend in any society toward increased permissiveness. There is some general evidence supporting this historical hypothesis in America.17
It may be argued that the twentieth century in America has witnessed precisely this sort of social change regarding sexual beliefs and behaviors. There is consistent evidence from past studies, such as those by Kinsey and Terman, to show trends toward increased permissiveness at all levels of sexual behavior.18 Chapters 7, 8, and 9 show how permissiveness changes in the various role stages of individuals. It is sufficient to say here that there appears to be arising in America a middle- and upper-class movement toward increased permissiveness, which seems strongly to resemble the Scandinavian sex patterns of the present day.19 In general, this can be seen in the vast popularity of such standards as "petting with affection," "transitional double standard" and "permissiveness with affection"-all of which accept individual-centered sexual behavior and beliefs. (See Exhibit 2.1 for definitions of all standards. )
In the past, it could be argued that groups with a tradition of high sexual permissiveness were groups that had less to "lose" by being highly permissive. Men and Negroes would be examples, for men cannot become pregnant and Negroes have less social standing to lose. Negroes have a high rate of divorce, desertion, and separation, which tends to reduce the attractiveness of a marital relationship as opposed to a nonmarital relationship. Among both Negroes and whites, the male role in the family institution is somewhat less emphasized socially than is the female role, and this may encourage a preference for the unmarried state and its benefits. Some support for this interpretation comes from a check on personal happiness in the national sample, which indicated that whites were happier than Negroes. Within the white group those who were married were the happiest. Within the Negro group those who were married were the least happy. This would indicate the lower relative status of the married state among Negroes and the strains associated with such a state. Generally, in both racial groups males were less happy than females. A control by social class showed that there was less of a difference between the married and single people in the lower and middle classes in terms of happiness than in the upper classes (the top one-third). However, in all race-sex groupings the basic relation of happiness to marriage, although weakened, was still present.
The current movement toward permissiveness among the middle and upper social classes seems to be based differently than it used to be. Rather than permissiveness being the philosophy of a group with less to "lose," it is a philosophy of a group with access to contraceptive controls on pregnancy, ways of combating venereal diseases, an intellectualized philosophy of the importance of ties based on affection and sex, and an integration of this viewpoint with a generally liberal position.20 The following chapters deal with this general view of sexual permissiveness in America and further test the specific proposition put forth concerning the relation of traditional levels of sexual permissiveness to social forces.21

----------------------

NOTES:

1 Some early comments on this area can be found in Ira L. Reiss, "Premarital Sexual Permissiveness Among Negroes and Whites," American Sociological Review, 29 (October 1964), pp. 688-698.
2 Among the many books on the American Negro, the following are most relevant to understanding Negro sexual behavior and beliefs: Harry S. Ashmore, An Epitaph for Dixie. New York: W. W. Norton & Co., 1957, and The Negro and the Schools. Chapel Hill: University of North Carolina Press, 1954; Jessie Bernard, Marriage and Family Among Negroes. New Jersey: Prentice-Hall. 1966; Brewton Berry, Race and Ethnic Relations. Boston: Houghton Mifflin Co., 1958; William Brink and Louis Harris, The Negro Revolution in America. N.Y.: Simon and Schuster, 1964; Allison Davis and John Dollard, Children of Bondage. New York: Harper Torchbooks, 1964; Allison Davis, Burleigh B. Gardner, and Mary R. Gardner, Deep South. Chicago: University of Chicago Press, 1941; John Dollard, Caste and Class in a Southern Town. New York: Doubleday Anchor, 1957; St. Clair Drake and Horace R. Cayton, Black Metropolis. 2 vols.; New York: Harper Torchbooks, 1962; E. Franklin Frazier, The Negro in the United States. New York: Crowell-Collier and Macmillan, 1957 and The Negro Family in the United States. New York: Holt, Rinehart and Winston, 1951; Paul H. Gebhard et al., Pregnancy, Birth, and Abortion. New York: Harper & Row, Publishers, 1958; Calvin C. Hernton, Sex and Racism in America. New York: Grove Press, 1965; Abram Kardiner and Lionel Ovesey, The Mark of Oppression. New York: Meridian Books, 1962; Albert J. Lott and Bernice E. Lott, Negro and White Youth. New York: Rinehart and Winston, 1963; Gunnar Myrdal, An American Dilemma. New York: Harper & Row, Publishers, 1944; John H. Rohrer and Munro S. Edmonson, eds., The Eighth Generation Grows Up. New York: Harper Torchbooks, 1960; George E. Simpson and J. Milton Yinger, Racial and Cultural Minorities. New York: Harper & Row, Publishers, 1953.
Few survey studies covering large regions of the country have been done, although some recent work of this kind is reported by Brink and Harris. Gebhard is an excellent empirical source for data from the Kinsey study comparing white and Negro females (see Chapter 6 of Pregnancy, Birth, and Abortion. )
3 All decks of IBM cards that were used in this study were checked in several ways. First, checks on coding errors were carried out. Marginals and listings on all decks were then made and checked. "Contradictory" punches-such as a male who listed his "husband's" occupation-were researched. It is believed that most errors were removed by this multiphase set of checks.
4 Leo Goodman, "Simple Methods for Analyzing Three-factor Interaction in Contingency Tables," Journal of the American Statistical Association, 59 (June 1964), pp. 319-352. Recently, Theodore R. Anderson supervised the programming of this test for the computer used in this study.
5 Alfred C. Kinsey et al., Sexual Behavior in the Human Female, Philadelphia: W. B. Saunders Co., 1953, pp. 304-307, 331. Kinsey found that religious devoutness was a better predictor of female than of male sexual behavior. See also Ernest W. Burgess and Paul Wallin, Engagement and Marriage. Philadelphia: J. B. Lippincott Co., 1953, p. 339; Rose K. Goldsen et al., What College Students Think. New York: D. Van Nostrand Co., 1960, p. 174; Eugene A. Kanin and David H. Howard, "Postmarital Consequences of Premarital Sex Adjustment," American Sociological Review, 23 (October 1958), pp. 556-562; and Winston W. Ehrmann, Premarital Dating Behavior. New York: Holt, Rinehart and Winston, 1959, p. 94.
6 The key measure of the strength of relationships used throughout this study is gamma. Q and gamma are the same for 2-by-2 tables, and since Q is better known, it is listed on the 2-by-2 tables. For a full discussion of gamma see three articles by Leo A. Goodman and William H. Kruskal, all published in the Journal of the American Statistical Association: "Measures of Association for Cross Classifications," 49 (December 1954), pp. 732-764; "Measures of Association for Cross Classifications 11: Further Discussion and Reference," 54 (March 1959), pp. 123-163; "Measures of Association for Cross Classification III: Approximate Sampling Theory," 58 (June 1963), pp. 310-364. C (coefficient of contingency) will be used in tables where the relation is clearly curvilinear. Note that since Q is a special use of gamma, it will often be referred to as gamma. For a good general discussion of the use of statistics see Leslie Kish, "Some Statistical Problems in Research Design," American Sociological Review, 24 (June 1959), pp. 328-338.
7 For the zero-order association, c2= 89.3 and Q = -.600. For all male students, c2=23.4 and Q = -.495; for all female students c2= 52.6 and Q = -.651.
8 The Y-square test was used for these measures. Goodman notes that this test is similar to a chi-square test, except it uses logs instead of frequencies.
9 The eight items in scale order are: (1) True love leads to almost perfect happiness; (2) When one is in love, the person whom he loves becomes the only goal in his life. One lives almost solely for the other; (3) True love will last forever; (4) There is only one real love for a person; (5) True love is known at once by the people involved; (6) Doubt may enter into real love; (7) Even though one's past love affair was not as strong as the present one, it may still have been a real love relationship; (8) Conflict can be a part of real love. Answers agreeing with the first five items and disagreeing with the last three produced the highest scale score on romantic love. The coefficient of reproducibility for the scale was .90 and the coefficient of scalability was .65. See Appendix A, Part VI for the full set of questions. Two questions were dropped for reasons discussed in Chapter 5.
10 The number of cases affects chi-square, and so one should always examine the gamma, or Q, rating in comparing tables. The test for interaction is also valuable as acomparison of tables of different size. Although the .05 level of significance will be used in this study the level of significance for .10 findings will be presented too, since in a one tail test these are actually at the .05 level. Zero cells in a 2 by 2 table affect Q. One zero cell makes Q = I and two zero cells makes Q = 0. For some warnings about pitfalls of tests for significance see Hanan C. Selvin, "A Critique of Tests of Significance in Survey Research," American Sociological Review, 22 (October 1957), pp. 519-527.
11 For recent empirical evidence see Ehrmann, chap. 4.
12 These measures are discussed in the next chapter.
13 The "Moynihan Report" was criticized for not using social-class controls to show how similar lower-class white and Negro families were. The evidence of the present study indicates that this similarity is illusory. Even with class controls the Negroes in the student sample came from larger families, with more fathers lower educated than mothers, and from more families with only a mother. For the Moynihan Report see The Negro Family, Office of Policy Planning and Research, United States Department of Labor, March 1965. For a review of the controversy see Lee Rainwater and William Yancy, "Black Families and the White House," Transaction (July-August 1966), pp. 6-11, 48-53. However, the reader should be aware of the difficulties involved in fully matching Negro and white groups. For some discussion of this see Jessie Bernard, "Marital Stability and Patterns of Status Variables,"Journal of Marriage and the Family, 28 (November 1966), pp. 421-429. See also the discussion of Bernard's article by Rainwater.
14 See Dollard, especially p. 249. Negro religious participation is higher than white; higher proportions of Negroes go to church, and there are proportionately more Negro churches. See Brink and Harris, chap. 6.
15 Although they are relatively permissive, Negroes in the sample for this study were not generally promiscuous. Table 2.6 shows that they tended to require affectionate relations as a basis for sexual behavior. Others have noted such a requirement (see Myrdal's classic study of the Negro, p. 935). As indicated in Chapter 2, Negroes were less inclined than whites to give high rank to affectionless kissing and petting. Some sex differences in both races are due to other factors, to be discussed in Chapter 5.
16 Christensen and Carpenter's work on three cultures also showed much smaller variation in male attitudes and behavior than in female attitudes and behavior. See their "Value-Behavior Discrepancies in Premarital Coitus," American Sociological Review, 27 (February 1962), pp. 66-74.
17 Ira L. Reiss, Premarital Sexual Standards in America, New York: The Free Press, 1960, chaps. 2 and 10 contain a large number of references and sources on this point.
18 Kinsey et al., Sexual Behavior in the Human Male and Sexual Behavior in the Human Female; Lewis Terman, Psychological Factors in Marital Happiness. New York: McGraw-Hill, 1938. For a recent statement see Ira L. Reiss, "The Sexual Renaissance, A Summary and Analysis," Journal of Social Issues, 22 (April 1966), pp. 123-137. These articles stress that many recent changes in coitus are in the area of attitudes rather than behavior. This is also discussed in chaps. 7 and 10 of the present study.
19 For a recent coverage of Scandinavian sex customs see Harold T. Christensen "Scandinavian and American Sex Norms," Journal of Social Issues, 22 (April 1966), pp. 60 - 75.
20 A regression analysis was performed on the tables in this chapter. No changes in interpretation occurred. In addition almost all the tables used in the study were examined by Somers "d" and Rosenberg's "standardization" approach. No changes in analysis or interpretation were required. See Robert H. Somers, "A New Asymetric Measure of Association for Ordinal Variables," American Sociological Review, 27 (December 1962), pp. 799-811, and Morris Rosenberg, "Test Factor Standardization as a Method of Interpretation," Social Forces, 41 (October 1962), pp. 53-61. The technique used to arrive at SEI scores was composed by Donald McTavish. The original deck referred to in his article has some minor errors, which were later removed. See Donald G. McTavish, "A Method for More Reliably Coding Detailed Occupations into Duncan's Socioeconomic Categories," American Sociological Review, 29 (June 1964), pp. 402-406.
21 Such relations as that between the number of times an individual has been in love and permissiveness could be viewed with permissiveness as the independent variable. However, the evidence is interpreted predominantly to support permissiveness as the dependent variable. This does not rule out the other possibility. Such interpretations are made from time to time on other relationships in order to present a particular causal point of view, but this should not be taken to mean the exclusion of all other causal patterns.