Trends in Teen Sexual Behavior
Douglas J. Besharov and Karen N. Gardiner
At least until the last few years, the proportion of teenagers who have had
sex rose steadily, and they are more likely to use contraceptives than their earlier
counterparts. Underlying this general trend has been an increase in sexual activity
among young teens—ages 15 and under—as well as middle-class and white
teenagers. Contraceptive use has risen also, probably because of the fear of AIDS
and the increased sexual activity among white and middle-class teenagers.
Because more teenagers were using contraceptives, the pregnancy rate per 1,000
sexually active teenagers actually declined during the 1980s, even as more
teenagers were having sex. Nevertheless, contraceptive use is far from perfect,
leading to high levels of abortion, sexually transmitted diseases, and nonmarital
births. The challenge for school-based programs is to pursue two simultaneous
goals: lower the level of sexual activity and raise the rate of contraceptive use
among those who have sex.
When President Clinton appointed his first surgeon general, the controversial and
outspoken Joycelyn Elders, he signaled his intention to make an issue of teenage
pregnancy and parenthood. Dr. Elders left Washington, but the issue has remained. The
past year, for example, witnessed the creation of a multimillion dollar national campaign to
reduce teenage pregnancy and welfare reform legislation that earmarked one-quarter
billion dollars for abstinence education.
There is nothing new about teenage sex. In the past, however, it was largely
confined to married teenagers, or those soon to be married, and thus was not considered a
problem. But today, teenagers are much less likely than in the past to get married. In
1993, for example, the average age at first marriage for women was 24.5 years and 26.5
years for men.
The decline in teenage marriages, however, has not led to a corresponding drop in
teenage sexual activity. During the 1970s and 1980s at least, more teenagers became
sexually active, and they had sex more frequently, with more partners, and at younger
ages. Newly-released data, however, indicate that there may have been a decline in
teenage sexual activity. According to the 1995 National Survey of Family Growth, the
proportion of sexually active females ages 15 to 19 fell 8% between 1988 and 1995
(Abma et al., 1997). The 1995 National Survey of Adolescent Males also found an 8%
decline in sexual activity among teenage males since 1988 (Sonenstein & Ku, 1997). This
is welcome news after the rate of sexual activity seemed to rise inexorably over the past
two decades. The new data contained another positive finding: The proportion of
sexually active teenagers using contraception rose dramatically.
These data may signal the start of a trend towards less sex and more contraceptive
use among teenagers. Further analysis of the data, as well as additional studies, will be
needed to draw a definitive conclusion. But even if these early suggestions of a
turnaround in teenage sexual activity are valid, there is still much room for improvement.
About half of male and female teenagers have had sex, including about 20% of 15-year-
old females. And contraceptive use is far from perfect. As a result, each year there are
about 1 million pregnancies resulting in approximately 400,000 abortions, 500,000 births,
and 100,000 miscarriages. Of the births, over 70% are out of wedlock, putting those
young mothers at risk of long-term welfare dependency. And about 3 million teenagers
contract a sexually transmitted disease (STD) such as chlamydia, gonorrhea, and even
Most people are unhappy about sexual activity among younger adolescents, but
disagreement is more widespread about sexual activity among older teenagers. (We will
not try to draw the age line.) Some believe that sex outside of marriage undermines
traditional moral values or can be psychologically damaging to adolescents. Others are
concerned that teenagers are relatively poor contraceptors.
Most agree that teenagers tend to be reckless in their behavior, and unprotected
sex is just one of a host of risky behaviors in which they engage. But even for those who
are always careful, accidents can happen and contraceptives can fail. Moreover, some
teenagers are exploited or coerced into having sex and thus have little control over
Today, several major policy issues surround teenage sexual activity: (1) how can
the level of such activity be reduced, especially among younger teenagers? and (2) how
can sexually active teenagers be better contraceptors? To provide a context for
considering these two questions, this paper presents an overview of major trends.
Five principal sources supply information about the sexual practices of teenagers: the
National Survey of Family Growth (NSFG), a national in-person survey of women ages 15
to 44 conducted in 1982, 1988, and 1995 (Forrest & Singh, 1990; CDC, 1991; Abma et
al., 1997); the National Survey of Adolescent Males (NSAM), a longitudinal survey of
males ages 15 to 19 conducted in 1988, 1991, and 19951 (Sonenstein, Pleck, & Ku, 1989
& 1991; Sonenstein & Ku, 1997); the National Survey of Young Men (NSYM), a 1979
survey of 17- to-19-year-olds (Sonenstein, Pleck, & Ku, 1989); and the Youth Risk
Behavior Survey (YRBS), a 1990 questionnaire-based survey of 11,631 males and females
in grades 9 to 12 conducted by the Centers for Disease Control (CDC, 1992). Additional
data are provided by the 1991 National Survey of Men (NSM), a nationally representative
sample of 3,300 males ages 20 to 39 (Billy et al., 1993).
All these surveys are based on self-reports and must be interpreted with caution.
Recall bias, for example, can be a problem. A 44-year-old may not remember the exact
circumstances surrounding her first sexual experience. Moreover, because many
interviews are conducted in person, the teenager or adult being interviewed may give the
answer that he thinks is socially correct. Thus, the following statistics should be viewed as
indicative of trends rather than as precise measures of current behavior.
More and Earlier Sex
Since 1970, nonmarital sexual activity among teenagers has increased sharply. The NSFG
asked women ages 15 to 44 to recall their first premarital sexual experience. About 40%
of the females who were 18 years old in 1970 said that they had had sex by that age. But
56% of the females who were 18 years old in 1980 said they had had sex by that age. For
those who were 18 in 1988, the proportion had risen to 70% (CDC, 1991). Premarital
teenage sex has always been more prevalent among males. According to the NSM, 55%
of males who were 18 years old in 1968–1970 had had sex, compared with 64% in 1977–
1979 and 73% in 1986–1988 (Alan Guttmacher Institute, 1994) (see chart 1).
CHART 1 HERE
As mentioned above, the 1995 NSFG reports a reversal in this trend. According
to the 1995 NSFG, the proportion of 18-year-old females who reported having had sex
fell 10%, to 63% (Abma et al., 1997).
Teenagers were not only more likely to become sexually active in the 1970s and
1980s, but they were experiencing sex at younger ages. As mentioned above, the
percentage of 18-year-old females who had had sex increased about 75% between 1970
and 1988. The percentage of sexually experienced 15-year-old females, however,
1As this paper was going to the press, data analysis from the 1995 National Survey of Family
Growth and the 1995 National Survey of Adolescent Males was limited. The data available to
us have been incorporated into this paper.
multiplied more than fivefold during the same period, from less than 5% to 26%. (See
The 1995 data, however, show that the proportion of teenage females who
reported that they have had sex fell or levelled off among all age groups (except 16-year-
olds). As noted above, sexual activity decreased about 10% among 18-year-olds (from
69% to 63%), stayed virtually the same among 17-year-olds (51% vs. 50%), and declined
almost 20% among 15-year-olds (from 26% to 21%) (Abma et al., 1997).2
CHART 2 HERE
Partly because they were beginning sex at earlier ages, the majority of teenagers in
the late 1980s reported having had sex with more than one partner (see chart 3).
According to the 1988 NSFG, the proportion of sexually active 15- to-19-year-old
females who had more than one lifetime partner increased from 38% in 1971 to 61% in
1988 (Smith, 1994). The 1990 YRBS found that about 7% of 9th-grade females and 19%
of males the same age had had intercourse with four or more different partners. By the
12th grade, 17% of females and 38% of males reported having had four or more sexual
CHART 3 HERE
The overall figures, however, tend to overstate teenage sexual activity because
they include all youngsters who have ever had sex, even once. Analyzing data from the
NSAM, for example, Sonenstein and her colleagues found that the average 15-to-19-year-
old male was abstinent for six months each year. Fifteen-year-old males spent an average
of eight months abstinent (Sonenstein, Pleck, & Ku, 1991).
Moreover, these are aggregate rates, and they mask important differences among
various groups in our society. Low-income and African-American teenagers have always
had higher rates of sexual activity than nonpoor and white ones, and still do. But the gap
2We are aware that a different study found that the increase in sexual activity among young
teenagers continued into the early 1990s. Using a different survey methodology, the CDC's
Youth Risk Behavior Survey (YRBS) found that in 1990, 32% of 9th-grade females (ages 14
and 15) reported ever having had sex, as did 49% of the males in the same grade. At the same
time, the proportion of 12th-grade females and males (ages 17 and 18) who reported ever
engaging in sex remained at about 1988 levels. Unlike the NSFG and NSAM, however,
which are face-to-face interviews, the YRBS is a paper-and-pencil questionnaire administered
to classrooms of students. No one knows whether this format leads to over- or underreporting
of high-risk behaviors. But the different methodology makes a comparison between these
figures and those from the NSFG and NSAM problematic.
is narrowing—at least among older teenagers.
Rates of sexual activity of poor and nonpoor females are converging, because
sexual activity has become more prevalent among middle-class and older teenagers.
According to the NSFG, between 1982 and 1988 the proportion of sexually active 15-to
19-year-old females in families with incomes equal to or greater than 200% of the poverty
line increased from 39% to 50%. At the same time, the proportion of females from poorer
families (those earning less than 200% of the poverty line) who had ever had sex remained
stable at 56% (Forrest & Singh, 1990) (see chart 4).
CHART 4 HERE
The differences in rates of sexual activity between white and black female
teenagers also narrowed during the 1980s. Analysis of the NSFG found that between
1970 and 1988, the percentage of 15- to 19-year-old black females who reported having
had sex rose from 49% to 59%. Among white females, the rate almost doubled between
1970 and 1988, from 27% to 51% (CDC, 1991). But these rates converged mostly
because older black and white females (ages 18 and 19) were having sex in similar
proportions: 76% and 73%, respectively (see chart 5).
CHART 5 HERE
Although more analyses will be needed to reach a definitive conclusion, it appears
that the black/white differential among females grew in the 1990s. According to the 1995
data, the proportion of black 15- to 19-year-old females who reported having had sex
stayed at the 1988 level. The proportion of sexually active white females, however, fell
slightly, to 47 percent (Abma et al., 1997). Even among older black and white females
(ages 18 and 19), the gap in sexual experience widened slightly. The proportion of blacks
who had had sex dropped one percentage point (to 75%), while the white rate fell 5
percentage points (to 68%).
Moreover, even before these recent data, it appeared that there were substantial
racial differences in sexual experience among younger teenagers (both boys and girls).
According to the 1988 NSFG, one-third of black 15-year-old females reported that they
had had sex, compared with 24% of their white counterparts (Moore, 1992).
Among males, the proportion of black 17-to-19-year-olds who had had sex
increased from 71% in 1979 to 88% in 1988, while the rate for white males increased from
64% to 73% (Sonenstein, Pleck, & Ku, 1989). Again, rates of sexual activity are closer
for older teenagers. The 1988 NSAM found that 96% of black males reported that they
had had sex by age 19 compared with 85% of white males. The gap was wider among 15-
year-olds: 64% of blacks versus 33% of whites reported already having had sex (Moore,
1992). The foregoing data measure the proportion of teenagers who have ever had sex.
Black teenager males, however, tend to report having had more partners and more
frequent sex. These racial differences also are highly associated with income differences,
so they should be interpreted with care.
More Careful Sex
Contrary to conventional wisdom, it appears that teenagers have become more careful
about contraception. The proportion of teenage females who reported using a method of
contraception at first intercourse, for example, increased from 48% in 1982 to 65% in
1988 to 77% in 1995 (Forrest & Singh, 1990; Abma et al., 1997). This is important
because studies indicate that half of premarital first pregnancies to teenagers occur within
six months of initiating sex, and one-sixth occur within a month of first sex (Zabin et al.,
Moreover, use of contraceptives at first sex increased among both white and black
teenagers. Although whites were still more likely to report having used a contraceptive,
the rate of increase since 1988 was greater among blacks. (see chart 6). The proportion of
white females who reported using a method of contraception at first intercourse increased
from 55% in 1982 to 69% in 1988. By 1995, use rose to 83%, a 20% increase. Among
black females, the proportion using contraception at first intercourse increased from 36%
to 54% between 1982 and 1988. By 1995, 72% of black females reported using a method
at first sex, a 33% increase from 1988 (Forrest & Singh, 1990; Abma et al., 1997).
CHART 6 HERE
Contraceptive use at first intercourse has also increased among males since the late
1980s. The 1995 NSAM found that proportion of 15- to 19-year-old males who said they
relied on an "effective" method of contraception (such as a condom or oral contraceptives)
increased from 62% in 1988 to 73% in 1995 (Sonenstein & Ku, 1997).
Much of the increase in contraceptive use at first intercourse was apparently driven
by condom use. The 1995 NSFG reported that 60% of females used a condom at first
intercourse, an increase from 47% in 1988. By way of comparison, the proportion of
females who relied on oral contraceptives increased from 8% to 11% (Forrest & Singh,
1990; Abma et al., 1997). Among males, condom use at first intercourse rose from 55%
in 1988 to 69% in 1995 (Sonenstein & Ku, 1997).
The 1995 findings concur with data from the 1980s that indicated the increased
use of contraception was apparently driven, at least in part, by fear of AIDS and other
sexually transmitted diseases. In the mid-1980s, a major public health campaign was
waged nationwide that promoted the use of barrier methods of contraception as a way of
avoiding STDs. Many teenagers heard this message. According to the NSAM, between
1979 and 1988, the proportion of teenage males who reported using condoms during their
last sexual experience almost tripled, from 21% to 57% (Sonenstein, Pleck, & Ku, 1989).
CHART 7 HERE
Increased condom use was also reflected in the NSFG data. The proportion of
female teenagers who reported that they relied on condoms (not just at last intercourse)
increased almost 75% between 1982 and 1988 (from 15% to 26%).
Increased contraception was also the result of more sexual activity among middle-
class and white youths, described above. These groups have traditionally been more
careful about their sexual practices. While contraceptive use at first sex increased among
all socioeconomic groups, teenagers from middle-class families were more likely to use
contraception at first intercourse. According to the NSFG, the proportion of middle-class
teenagers who reported using contraception at first sex increased between 1982 and 1988,
from 57% to 73%. During the same period, use at first intercourse among teenagers from
poorer families increased from 40% to 58% (Forrest & Singh, 1990).
CHART 8 HERE
Because teenagers are more likely to use birth control, the substantial increase in
teenage sexual activity has not been matched with an equivalent increase in the number of
teenage conceptions. In fact, the absolute number of pregnancies increased by about 21%
between 1975 and 1988, even as there was a 39% increase in the number of female
teenagers who reported engaging in sex. The pregnancy rate among sexually active
teenagers (as opposed to the entire population of teenagers) actually declined between
1982 and 1987, from 271 per 1,000 teenagers to 243 per 1,000 (Forrest & Singh, 1990).
In fact, one could construct a crude "teen conception index" that measures the
changing rate of conception (composed of abortions, miscarriages, and births) among
sexually active but unmarried teenagers. This index stood at 0.87 in 1988, representing a
13% decline from 1975. Most of this decline occurred between 1985 and 1988 as
contraceptive use increased among all groups and as more middle-class teenagers initiated
sex (Besharov & Gardiner, 1993)3 (see chart 8). If, as the 1995 data indicate, levels of
3 Conception index sources: for abortions, Stanley K. Henshaw, the Alan
Guttmacher Institute, based on 97% of abortions to unmarried teens. For miscarriages:
Stanley K. Henshaw, the Alan Guttmacher Institute, assuming that all were to unmarried
teens. For out-of-wedlock births: Stephanie Ventura, National Center for Health
Statistics. For the total number of unmarried females ages 15–19: the Census Bureau,
Division of Marriage and Family. For the percentage of unmarried, sexually active
females ages 15–19: the National Center for Health Statistics, National Survey of
teenage contraceptive use rose, we would expect the index to drop even further.
Contraceptive Nonuse and Misuse
Despite the improvements in contraceptive use, however, many teenagers still use
contraceptives inconsistently, or not at all. Although contraceptive use at first intercourse
did increase during the 1980s, the NSFG reports that in 1988, 31% of white female
teenagers and 41% of black ones reported that they did not use any method at that time.
Moreover, in that year, 20% o sexually active female teenagers reported that they were
not using any form of contraception (Forrest & Singh, 1990). Contraceptive use at first
intercourse continued to increase during the 1990s. The proportion of teenage females
engaging in unprotected first sex fell sharply, to 17% of whites and 28% of blacks. There
is room for substantial progress on this front, however. Almost one in five white females
and one in three black females were not using birth control during their first sexual
experience, thus putting themselves at risk of pregnancy and disease (Abma et al., 1997).
The more risky the teenager's general behavior is, the greater the likelihood is of risky—
and early—sex (see chart 9). Donald Orr and his colleagues at the University of Indiana
studied condom use among inner-city female teenagers in Indianapolis and found that only
about 50% of them reported that their partners had ever used condoms. Significantly, the
figure fell to only 29% among "risk-takers," defined as drug or alcohol users, school
dropouts, and those who committed minor delinquencies (Orr et al., 1992).
CHART 9 HERE
As chart 10 indicates, sex among young teenagers is also associated with other
risk-taking behaviors such as smoking, drinking, and using drugs. The YRBS found, for
example, that the proportion of 14- and 15-year-olds who had had sex was 87% among
those who smoked marijuana regularly, compared with 36% of those who never did so
CHART 10 HERE
Moreover, many teenagers report that they were using a contraceptive at the time
they got pregnant but that it failed. As chart 11 shows, contraceptive failure rates within
the first year of use are higher for teenagers, particularly poor ones, than among "typical
users" in the general population. Contraceptive failure rates are often presented in two
For conception index calculations: For the base year (1975), total pregnancies were
divided by total number sexually active. For subsequent years, total pregnancies were
divided by total number sexually active for each year, then divided by the base year
ways: (1) failure rates in clinical trials (usually involving married couples); and (2) the
failure rates of "typical users" (based on surveys of the general population). James
Trussell of Princeton University estimates, for example, that the clinical failure rate of the
condom is 2%. Twelve percent of typical users, however, experience a condom failure in
their first year of use, compared with 13% of never-married, middle-class teens and 27%
of poor ones. The difference in failure rates is even greater for less effective methods.
About one in five typical users will experience a failure while relying on periodic
abstinence. The rates for never-married, middle-class teenagers and poor teenagers are
27% and 52%, respectively (Trussell et al., 1990; Jones & Forrest, 1992).
CHART 11 HERE
Coerced or Exploitative Sex
Another obstacle to both abstinence and safer sex is the apparently high level of coerced
and exploitative sex. Kristin Moore of Child Trends analyzed data from the 1987
National Survey of Children and found that three-quarters of the females who had had
intercourse before age 14 and 60% of those who had had sex before age 15 reported
having had sex involuntarily (AGI, 1994). The National Health and Social Life Survey
also tried to gauge the "wantedness" of first sex. Fewer women than men reported that
their first sexual experience was wanted (71% vs. 92%) (Laumann et al., 1994).
In recent years, there has been growing concern among policy makers and the
public that a large number of births to teenage mothers–50% by some accounts—involve a
father who is not himself a teenager. This belief has led to a renewed interest in statutory
rape laws. Society should certainly be concerned about exploitative relationships. But the
50% figure is misleading because it includes many parents who are close in age but happen
to fall on opposite sides of the 20-year mark. In other words, the bulk of these situations
involve mothers in their late teens and fathers in their early-to mid-20s.
Laura Lindberg and her colleagues at the Urban Institute give us a more precise
picture of what is happening. Using the 1988 National Maternal and Infant Health Survey,
they estimated the number of births to unmarried minors (ages 15 to 17) that involved men
five or more years older. (They chose a five-year difference because that is the measure of
statutory rape in many states.) The authors found that 8% of births to 15- to-19-year-olds
involved unmarried minors and older males. Only 3% of all nonmarital births in 1988 did
so (Lindberg et al., 1997).
The authors did find, however, that the younger the mother, the more likely that
the relationship involved an older man. About 40% of 15-year-old mothers had a partner
five or more years older, compared with 27% of 16-year-olds and 24% of 17-year-olds.
(Births to 15-year-olds, however, represent only 13% of all births to 15-to-17-year-olds.)
These findings are corroborated by birth records. Mike Males of the University of
California at Irvine studied 1990 California birth certificates and found that when the
mothers were ages 11 to 12, the fathers of their babies were, on average, 10 years older.
Among mothers ages 13 to 14, the fathers averaged 4.6 years older (Males, 1993) (see
CHART 12 HERE
The costs to society of unprotected sex are considerable. In 1992, for example, there
were about 1 million pregnancies, resulting in over 300,000 abortions, 134,000
miscarriages, and 500,000 live births (Henshaw, 1992). Of the births, 70% were out of
wedlock (National Center for Health Statistics, 1994). And about 3 million teenagers
suffered from a sexually transmitted disease such as chlamydia, syphilis, gonorrhea,
herpes, and even AIDS (AGI, 1994).
Abortion: The Abortion Provider Survey, performed periodically by AGI, collects
information about abortions and the clinics that provide them. According to AGI, about
35% of all teenage pregnancies now end in abortion and another 14% in miscarriage.
(Unmarried teenagers account for about 97% of all abortions to teens.) Thus, of the 1.5
million abortions in 1992, over 300,000 were performed on teenagers. In the 11 years
between 1973 and 1984, the abortion rate for females ages 15 to 19 almost doubled, from
23 to 43 per 1,000 (AGI, 1994). Between 1984 and 1992, the rate stabilized.
CHART 13 HERE
Chart 13 depicts the proportion of abortions by income subgroup. Poor teenagers
had about the same proportion of abortions (32%) as middle-income teenagers. (These
numbers reflect the distribution for all women.) These figures, however, are misleading
because middle-income teenagers account for a larger proportion of the general
population. In 1987, AGI calculated an abortion index by dividing the proportion of
abortion patients in a subgroup by the proportion of women ages 15 to 44 in the same
subgroup. (Estimates were not available for teenagers.) An index of 1.0 reflected the
national average. The researchers found that the index for poor women stood at 2.22,
meaning that they had abortions at a rate more than twice the national average. The index
for middle-class women, 0.59, indicated that they had abortions at about half the national
rate (Henshaw and Silverman, 1988).
Nonmarital births: According to the National Center for Health Statistics, more
than 350,000 babies were born to unmarried 15-to-19-year-olds in 1992. Although the