Divorce and Sex

Abstract

Studies on the sex lives of the divorced are relatively few and somewhat dated. Although they give a rather optimistic view of the subject, this research is flawed by high respondent refusal rates and poor representativeness of samples. Further, a major gap in this literature is that the possible predictors of sexual activity of the divorced remain largely unexplored. The present study, using national data on 340 divorced people, indicates a much lower level of sexual activity than found in past research. Regression analyses find that both number of partners and sex frequency are related to religiosity, education, and political liberality. Future research can explore the impact of low sexual activity among the divorced on indicators of emotional well-being.

Full text

Archives of Sexual Behavior, Vol. 21, No. 4, 1992
Divorce and Sex
Steven Stack, Ph.D. 1,3 and James H. Gundlach, Ph.D. 2
Studies on the sex lives of the divorced are relatively few and somewhat dated.
Although they give a rather optimistic view of the subject, this research is flawed
by high respondent refusal rates and poor representativeness of samples.
Further, a major gap in this literature is that the possible predictors of sexual
activity of the divorced remain largely unexplored. The present study, using
national data on 340 divorced people, indicates a much lower level of sexual
activity than found in past research. Regression analyses find that both number
of partners and sex frequency are related to religiosity, education, and political
liberality. Future research can explore the impact of low sexual activity among
the divorced on indicators of emotional well-being.
KEY WORDS: divorce; age; gender; sex.
INTRODUCTION
Sexual activity has been associated with indicators of individual well-
being (Allgeier and Allgeier, 1988). These include depression (Beck, 1968;
Favan et al., 1982), longevity and vigor (Verwoerdt, 1976), affective disor-
ders (Schreiner-Engel and Schiavi, 1986), educational attainment (Bourque
and Cosand, 1989), insomnia (Butler and Lewis, 1973), suicide (Leviton,
1973), and as an enhancement of the quality of life of recovering patients
(Gordon, 1986; Turner et al. 1989). Unfortunately, no systematic work has
been done on the link between sex and well-being for the divorced popu-
lation per se. The well-known emotional problems of living among the di-
The Interuniversity Consortium for
Political and
Social Research provided the datatape
containing the data used in this paper.
1Department of Sociology, Wayne State University, Detroit, Michigan 48202.
2Auburn University, Auburn, Alabama 36849.
3To whom correspondence should be addressed.
359
0004-0002/92/0800-0359506.50/0 © 1992 Plenum Publishing Corporation

360 Stack and Gundlach
vorced, such as depression (e.g., Gerstel
et al.,
1985) and suicide (e.g.,
Stack, 1990), might be associated with a low level of sexual activity, but
this is largely undocumented.
Research on the sexual behavior of the divorced is sparse. Neverthe-
less, existing generalizations often give the impression that the sex lives of
divorced people are about the same (and in some instances better) as those
of married people. From this, however, we would not anticipate the sub-
stantial high level of problems such as depression and suicide among the
divorced. Such problems would have to be totally a result of nonsexual
conditions such as financial worries.
Hunt (1974) is typically cited as the authority on the extent of sexual
activity among the divorced. His study, however, is flawed on methodologi-
cal grounds. The present study corrects these methodological problems and
reexplores the sexual behavior of the divorced. It provides epidemiological
estimates of sexual activity using national data. Given that these are in fact
much lower, they have implications for understanding the problems of living
among the divorced. It then runs regression models to determine what so-
cial variables are the best predictors of sexual activity among the divorced.
These might be used in understanding what subgroups of the divorced are
most at risk for low sexual activity and associated problems of living.
Previous Work
The sexual behavior of the divorced population has received minimal
attention. Butler (1979) argued that so little is known that the subject has
been one marked by "wild speculation." This blind spot in the literature
is significant given that the divorced population is growing and currently
constitutes 9% of the adult population (U.S. Bureau of the Census, 1989).
Other than the brief treatment of sex among the divorced found in
the classic Kinsey
et al.
(1948, 1953) studies, the only other national study
is that by Hunt (1974). His research is based on 2026 interviews with mar-
ried, single, divorced, and widowed people However, 80% of the respon-
dents refused to participate. This high rejection rate makes generalizing
the results to the divorced population questionable.
Hunt dealt with five issues on sexuality and the divorced. Two can
be dealt with using our data: the number of partners in the past year and
the frequency of sexual intercourse. With respect to number of partners,
for men the range was from a minimum of one to several dozen, and the
median was eight. For women the median was four partners per year. With
respect to frequency, men reported having sex somewhat more than twice
a week (double the amount found by Kinsey). Women reported having sex

Divorce and Sex 361
slightly less than twice a week, on average (four times the frequency found
by Kinsey).
The present study updates these two issues employing data from the
recent General Social Surveys (Davis, 1989). In addition, it goes beyond
Hunt's selection of two independent variables and bivariate mode of analy-
sis. It also employs multivariate techniques using a series of independent
variables drawn out of the more general literature on sexual behavior.
We first test the following hypotheses: drawn from Hunt's work: (i)
Divorced males will be more sexually active than divorced females and (ii)
younger divorced persons will be more sexually active than older persons.
In addition, we test hypotheses drawn from the more general literature.
First, religiosity has generally been found to lower the incidence of sex
outside of marriage, since religions tend to teach against such sexual pat-
terns. The more persons are tied to religion, the less likely they are to
deviate from sex outside of marriage. Research has found religiosity in-
versely related to extramarital sex attitudes (Reiss
et al.,
1980; Weis and
Jurich, 1985). Research also finds an inverse relationship between religios-
ity and sex before first marriage (e.g., Reiss and Miller, 1979; Miller and
Bingham, 1989). Finally, religiosity (as church attendance) has been found
to act as a deterrent to cohabitation (e.g., Tanfer, 1987). Although no work
has rigorously tested the religious perspective on sexuality among the di-
vorced, we contend that it might be generalizable to that population: The
greater the religiosity, the lower the number of sex partners and sex fre-
quency.
Consistent with previous research (e.g., Weis and Jurich, 1985; Reiss
et al.,
1980; Tanfer, 1987; Davidson and Leslie, 1976) we also introduce
independent variables measuring education level, political liberality, com-
munity size, living in the far West, and the absence of children (all thought
to be positively related to sexual activity). The absence of children is taken
as an opportunity factor for sex as the absence of parents is in the literature
on premarital sex (e.g., Davidson and Leslie, 1976).
METHOD
Data are from the 1988 and 1989 General Social Survey (GSS). These
surveys are full probability samples of noninstitutionalized, English-speak-
ing persons 18 years of age or older, living in the U.S. Subjects are inter-
viewed in their homes by trained interviewers. The 1988 response rate was
77.3% and the 1989 rate was 77.6% (Davis, 1989). For a further discussion
of sampling techniques, demographic characteristics of the sample, and re-

362 Stack and Gundlach
lated issues see Davis. An advantage of the GSS is that it is based on a
national random sample. There are 340 divorced respondents.
Two items on sex frequency are available in the GSS surveys: the
number of sex partners the respondent has had in the past year, and the
frequency (e.g., never, 1-2 per year, once a month) of sexual intercourse.
Both surveys have the question on number of sex partners; but the 1989
survey was the first to have the question on sex frequency. Number of sex
partners was collapsed into three categories given relatively few in the high
response categories of the item. The categories for the bivariate analysis
are none, one, two or more partners.
Since ordinary least squares (OLS) regression techniques may not be
robust enough to handle a trichotomous dependent variable, the study also
collapses the dependent variable into two categories for a logistic regression
analysis: no sex partners, and one or more sex partners (Morgan and
Teachman, 1988).
Gender is a binary variable (coded 0 for males). Preliminary analysis
of the data indicated a split in sexual activity between people under vs over
35. Age is, therefore, coded as 0 for less than 35, and 1 for 35 and over.
Religiosity is measured in terms of church attendance as coded in the GSS.
As in the past research, education is measured as the number of years
completed. Political liberality is measured as in Reiss
et al.
(1980): a seven-
point, self-rated scale from extremely conservative to extremely liberal.
Community size is measured following the conventions in Weis and Jurich
(1985): a six-point scale from rural counties with no town as large as 10,000
through a high of the 12 largest central cities. Alternative measures such
as residing in a place of over 100,000, however, yielded the same results.
The western region variable is treated as a binary variable (0 = nonwest,
1 = west). Absence of children is also treated as a binary variable (1 =
no children in the household).
It would be desirable to take into account which, if any, of the di-
vorced were cohabiting. This is not possible, however, with the GSS data.
The closest item is unrelated persons in the household, but their gender
is not given. We cannot determine if it is an unrelated person of the op-
posite sex.
Analysis
Table I gives the distribution of the number of sex partners by gender
and age. For all divorced persons, 74% had either no partners or one part-
ner. The median number of partners is less than one, less than a quarter
of the amount found in previous work. As in past research, gender is as-

Divorce and Sex 363
sociated with the number of sex partners. Sixteen percent of men had no
sex partners compared to 34% of women. Divorced men are more sexually
active than divorced women. Age is inversely related to sex. Only 4% of
those under 35 had no sex partners compared to 35% of those over 35.
Thus, both hypotheses drawn from Hunt's work concerning age and gender
are supported.
Table II provides data on sex frequency. These indicate a lower level
of sexual activity than in past work. The median is now twice a month as
opposed to twice a week.
Tables III and 1V provides the results of the multivariate analysis of
the number of sex partners. Table III employs OLS techniques with the
trichotomous dependent variable. Controlling for the other independent
variables, both age and gender remain strong predictors of sexual activity.
A second significant predictor is religiosity. The lower the religiosity the
higher the probability of having a sex partner. Education and political lib-
erality are also significantly related to sexual activity. The beta coefficients
indicate that age and gender are the most important predictors of the num-
ber of sex partners. Community size, the absence of children, and residing
in the West are not, however, related to sexual activity. The model explains
21% of the variance. Turning to Table IV, the results of a logistic regression
on the dichotomous dependent variable are essentially the same. The OLS
estimates are largely robust. The only exception is that absence of children
is now also significant. The logistic analysis correctly classified 78% of the
respondents as inactive or active sexually.
Table 1. Associations Between Number of Sex Partners, and Gender and Age, 1988-1989
No. of Total Male Female < 35 >_ 35
partners (N = 340) (n = 109) (n = 231)
(n
= 74) (n = 266)
None 27.9 15.6 33,8 4.1 34.6
One 45.9 41.3 48,1 58.1 42.5
Two or more 26.2 43.1 18,2 37.8 22.9
Table 11. Distribution of Frequency of Sexual Intercourse, 1989 only (n = 158)
Frequency %
Not at all 25.3
Once or twice/year 8.9
Once a month 7.0
2-3 times/month 15.8
Weekly 17.7
2-3 times/week 19.6
4+ per week 5.7

364 Stack and Gundlach
Table III. The Effect of Gender, Age, and Other Variables on the Number of Sex
Partners (0, 1, 2+) Among the Divorced, 1988-1989
(n
= 323), Ordinary Least Squares
Estimates
Variable Coefficient SE t 13
Intercept 2.17 0.27 7.92 a --
Church attendance -0.31 0.02 -2.05 a -0.11
Age -0.46 0.91 -4.97 a -0.25
Gender -0.43 0.08 -5.16 a -0.27
Urban residence 0.11 0.09 1.18 0.06
Western residence -0.14 0.09 -1.63 -0.08
Education 0.03 0.01 2.12 a 0.11
Absence of children -0.10 0.07 -1.26 -0.07
Political liberality 0.08 0.03 2.74 a 0.14
R 2 0.21
ap < 0.05.
Table IV. The Effect of Gender, Age, and Other Variables on the Presence or Absence of
Sex Partners Among the Divorced, 1988-1989 (N = 340), Logistic Regression °
Variable 13 SE )~2
Intercept 5.26 1.20 19.38 b
Church attendance -0.10 0.06 2.94 b
Age -2.46 0.62 15.88 b
Gender - 1.33 0.35 14.17 b
Urban residence 0.44 0.37 1.47
Western residence -0.47 0.33 1.93
Education 0.15 0.05 8.31 b
Absence of children -0.83 0.30 7.91 b
Political liberality -0.27 0.11 5.64 b
aNumber of cases correctly classified: 78.4%.
bp
< 0.05.
Table V provides the OLS estimates for predicting the frequency of
sexual intercourse. Only gender and age emerge as significant predictors of
sex frequency. Males and the young are apt to have sex more frequently than
females and older people. Other independent variables were not significantly
related to sex frequency. The smaller N (1989 sample only) covered by Table
V may account for this; the model explains 22% of the variance.
CONCLUSION
Results indicate that the estimates of sex activity in the past research
by Hunt (1974) and Kinsey
et al.
(1948, 1953) are high. The findings of

Divorce and Sex 365
Table V. The Effect of Gender, Age, and Other Variables on the Frequency of Sexual
Intercourse 1989, Ordinary Least Squares Estimates (n = 146)
Variable Coefficient SE t [3
Intercept 6.41 1.18 5.40 a --
Church attendance -0.003 0.06 -0.04 0.004
Age - 1.40 0.38 -3.71 a -0.29
Gender - 1.54 0.36 -4.34 a -0.35
Urban residence 0.45 0.40 1.14 0.09
Western residence -0.25 0.36 -0.70 -0.05
Education 0.03 0.06 0.51 0.04
Absence of children -0.31 0.33 -0.95 -0.08
Political liberality 0.07 0.13 0.55 -0.04
R e 0.22
ap
< 0.05.
the past work, where a majority of the respondents elected not to partici-
pate, need to be taken as tentative. In contrast, the present study is based
on just a few questions on sexuality which were sandwiched between liter-
ally hundreds of others on a variety of topics. Hence, the cooperation of
the respondents was much greater.
With a national random sample of divorced persons, we find that the
median number of sex partners and sex frequency are about only one quar-
ter of what the earlier work estimates. We speculate that the Kinsey and
Hunt studies elicited a differentially high response rate among the more
sexually active divorced population.
Age, gender, religiosity, education, political liberality, and the op-
portunity factor (absence of children) were able to predict correctly the
number of sex partners of 78% of the respondents. Past work has often
found evidence linking the other independent variables to aspects of
sexuality. Some of this research is, however, concerned only with attitudes
towards sexuality as its dependent variable (e.g. Reiss
et al.,
1980; Weis
and Jurich, 1985). Possibly some variables found to predict attitudes will
not predict actual behavior. Some divorcees may, for example, have posi-
tive attitudes about sexual liaisons with attractive members of the oppo-
site sex, but may be unable to find partners willing to meet their
expectations.
The variables in the model were largely unrelated to sex frequency.
Two exceptions were age and gender. The full range of variables worked
better in predicting the number of sex partners than in predicting the num-
ber of times one engages in sex with available partners. As with the case
of predicting number of partners, age and gender were the variables most
closely tied to sexual frequency.

366 Stack and Gundlach
Given that we find a much lower incidence of sexual behavior among
the divorced than was previously thought to be the case, this may be a
particularly robust point of departure for studies of well-being, including
work on depression and suicide. Unfortunately, this issue cannot be ad-
dressed adequately with the GSS data. Such studies on well-being, however,
tend to routinely leave out sexuality as a possible contributing factor. Re-
ports from clinicians, however, tend to list sexual tensions among the di-
vorced as a leading contributor to negative psychological states such as
anxiety, depression, and low self-esteem (e.g., Fisher, 1987).
In any event, our new, lower estimates of sex among the divorced
are more consistent with the literature on the effects of divorce on well-
being than the old, relatively high estimates of sexual activity which implied
equal levels of sexuality between the divorced and married. These more
accurate estimates of sexuality might be used by clinicians working with
divorced people in communicating realistic expectations for sexuality. The
findings from the multivariate analysis might also help therapists in isolating
groups at high risk for sexual tensions.
REFERENCE
Allgeier, A. R., and Allgeier, E. R. (1988). Sexual Interactions, D.C. Heath, Lexington, MA,
pp. 475-476.
Beck, A. (1968). Sexuality and depression. Med. Aspects Hum. Sex. 2: 44-51.
Bourque, L., and Cosand, B. (1989). Predicting the educational attainment
of Los
Angeles
women. Social. Perspect. 32: 137-151.
Butler, E. (1979). Traditional Man'iage and Emelging Alternatives, Harper and Row, New York.
Butler, R. N., and Lewis, M. I. (1973). Aging and Mental Health, C. V. Mosby, St. Louis, MO.
Davidson, J. K., and Leslie, G. R. (1976). Premarital sexual intercourse: An application
of
axiomatic
theory construction.
J. Mar. Faro. 39: 15-25.
Davis, J. A. (1989). General Social Surveys, 1972-1989: Cumulative Codebook, National
Opinion Research Center, Chicago, IL.
Favan, G., Muniari, F., and Pavan, L. (1982). Depression and sexuality: A controlled study.
Riv. Sper. Freniatr. 106: 1259-1264.
Fisher, B. (1987). Rebuilding: When Your Relationship Ends, Impact Publishers, San Luis
Obispo, CA.
Gerstel, N., Riessmann, C., and Rosenfield, S. (1985). Explaining the symptomatology
of
separated and divorced women and men. Social Forces 64: 84-101.
Golden, J. (1986). Sexual problems of cancer patients. Med. Aspects Hum. Sex. 20: 70-76.
Hunt, M. (1974). Sexual Behavior in the 1970's, Dell, New York.
Kinsey, A. C., Pomeroy, W., and Martin, C. (1948). Sexual Behavior in the Human Male, W.
B. Saunders, Philadelphia, PA.
Kinsey, A. C., Pomeroy, W., Martin, C., and Gebhard, P. (1953). Sexual Behavior in the Human
Female, W. B. Saunders, Philadelphia, PA.
Leviton, D. (1973). The
significance of sexual activity
as a determinant
to suicide among the
aged. Omega 4: 163-174.
Miller, B., and Bingham, C. R. (1989). Family configuration in relation
to the sexual behavior
of
female adolescents. J. Marr. Faro. 51: 499-506.

Divorce and Sex 367
Morgan, S. P. and Teachman, J. (1988). Logistic regression: Description, examples, and
comparisons. J.
Marr. Fam.
50: 929-936.
Reiss, I., and Miller, B. C. (1979). Heterosexual permissiveness a theoretical analysis. In Burr,
W., Hill, R., Nye, F. I., and Reiss, I. (eds).,
Contemporary Theories About the Faro@,
Vol. 1, Free Press, New York.
Reiss, I., Anderson, R. E., and Sponaugle, G. C. (1980). A multivariate model of the
determinants of extramarital sexual permissiveness. J.
Marr. Farn.
42: 395-411.
Schreiner-Engel, P., and Schiavi, P. (1986). Lifetime psychopathologies in individuals with low
sexual desire. J.
Nerv. Ment. Dis.
174: 646-651.
Stack, S, (1990). New micro-level data on the impact of divorce on suicide, 1959-1980: A test
of two theories.
J. Marr. Fam.
52: 119-127.
Tanfer, K. (1987). Patterns of premarital cohabitation among never married women in the
United States. J.
Mar. Fam.
49: 483-498.
Turner, L., Althof, S., and Levine, S. (1989). Self injection of papaverine and phentolamine
in the treatment psychogenic impotence.
J. Sex Marital Ther.
15: 163-176.
U.S. Bureau of the Census. (1989).
Statistical Abstract of the United States, 1989,
U.S.
Government Printing Office, Washington, DC.
Verwoerdt, A. (1976).
Clinical Geropsychiatry,
Williams and Wilkins, Baltimore, MD.
Weis, D., and Jurich, J. (1985). Size of community of residence as a predictor of attitudes
towards extramarital sex. J.
Marr. Fam.
47: 173-183.