An overwhelming majority of U.S. adults have long supported sexuality education in the public schools, according to a wide array of surveys.¹ That support extends not only to teaching about abstinence,* but also to teaching about contraception for the prevention of pregnancy and of sexually transmitted diseases (STDs), including HIV. Moreover, according to a 1997 survey, while eight in 10 adults believe it is very important that teenagers be given a strong message from society that they should abstain from sex until they are at least out of high school, six in 10 say that sexually active young people should have access to birth control, and only two in 10 object to that proposition.²

Public opinion regarding the scope of sexuality education is consonant with the weight of research in this area. In February 1997, for example, the Consensus Panel on AIDS of the National Institutes of Health declared that the "abstinence-only" approach "places policy in direct conflict with science because it ignores overwhelming evidence that other programs are effective" in delaying the onset of sexual intercourse among adolescents and in reducing their number of partners and increasing their condom use, if they are already sexually active.³

Indeed, a 1997 comprehensive review and analysis of existing evaluations of abstinence-only programs concluded that while there may be too little evidence for a definitive conclusion, "there does not exist any scientifically credible, published research demonstrating that they have actually delayed . . . the onset of sexual intercourse or reduced any other measure of sexual activity."⁴ That analysis, which also assessed other approaches to sexuality education and teenage pregnancy prevention, supports the major conclusions of international literature reviews conducted in 1993 and 1997—that the programs most effective in changing young people's behavior, in terms of both delaying their initiation of sexual intercourse and promoting their eventual contraceptive use, are those that address abstinence along with contraception for pregnancy and STD prevention (often termed a "comprehensive" approach).⁵

Nonetheless, among U.S. policymakers at the federal and state levels, educational efforts that focus narrowly or exclusively on abstinence promotion are being widely embraced. In 1996, as part of comprehensive welfare reform legislation, Congress established a new $250 million, five-year entitlement to states to support a variety of educational efforts, including but not limited to school-based programs, that must have abstinence promotion outside of marriage as their "exclusive purpose." These efforts must also be entirely separate from state programs that involve contraceptive information or services.⁶ To date, all but two states have accepted federal funds under these conditions and are currently in various stages of implementing their programs.⁷

The enactment of this 1996 federal law was a milestone in a concerted effort over the past two decades by self-described "profamily" groups to advocate the promotion of abstinence, rather than contraceptive education and services, as the appropriate strategy for addressing teenage sexual activity and pregnancy. As far back as 1981, Congress enacted the Adolescent Family Life Act (AFLA), which among other goals sought to establish a counseling and service network parallel to the Title X-funded family planning clinic system that would promote "self-discipline and other prudent approaches to the problem of adolescent premarital sexual relations." Such a system also would have provided contraceptive services only to adolescents who already had a child and who were seeking to prevent a second or subsequent birth.⁸ While this national service network never materialized, the AFLA subsidized the development of curricula that became a central organizing tool for promoting abstinence-only education in schools.

The Sexuality Information and Education Council of the United States (SIECUS) has extensively, if anecdotally, documented local controversies surrounding sexuality education since the 1991-1992 school year. By 1995, SIECUS had tracked more than 400 controversies over the preceding three years; it identified 160 new conflicts that surfaced in 40 states during the 1994-1995 school year alone.⁹ By the 1996-1997 school year, the cumulative total had risen to more than 500 local controversies in all 50 states. Most of these controversies involved groups promoting abstinence-only programs over the existing or proposed sexuality education program.¹⁰ Still, relatively little is known about the impact on school policies of this lengthy and ongoing campaign.

In this article, we present results from the first nationwide assessment of the extent to which sexuality education policy at the local school district level has focused on the promotion of abstinence. We pay specific attention to the relationship between policy on teaching abstinence and policy on providing contraceptive information. We examine existing policies nationwide and how they vary by district size, metropolitan status and region. We also explore school superintendents' perceptions of the factors that most influenced how their policies were established.

Data and Methods

Sample

The sample frame for the analysis comes from the U.S. Department of Education's National Center for Education Statistics. We used the early release of the Common Core of Data, Public Elementary/Secondary Education Agency Universe for the school year 1996-1997.¹¹ These data contain the names of all public school districts in the United States, their mailing addresses, the grades taught and geographic indicators, such as metropolitan status. We combined files for the 50 states and the District of Columbia to create a database of all 16,448 public education districts. We excluded from this overall total the 1,537 administrative districts that had no students enrolled.

Since our survey was designed to collect information on policies in school districts that taught grade six or higher, we deleted a further 1,346 school districts that only included grade five or for which grade-level information was unavailable. Of the districts initially sampled, five were later found to be ineligible because they had closed or were duplicated by another case; thus, the corrected sample frame contained 13,560 eligible school districts.

We stratified the districts by numbers of students so we could compare policies according to enrollment size; these groups were 1-4,999 students (small), 5,000-24,999 students (medium) and 25,000 students or more (large). Within these strata, we sampled all 224 large-enrollment districts, and we randomly sampled 500 districts in both the small- and medium-enrollment districts, for a total of 1,224 sampled districts.

Fielding

Questionnaires addressed to "Superintendent" were mailed to each sampled district in late May 1998, with reminder postcards sent one week later. We called nonresponding school districts beginning in late June to verify addresses and to obtain the name of the superintendent. We then sent a second questionnaire, with a cover letter addressed to the superintendent by name. Districts that still had not responded were called a second time, and the interviewer attempted to speak with the superintendent or a person to whom the superintendent might delegate responsibility for completing the questionnaire. A third questionnaire was then mailed or faxed to the person identified as most likely to complete the questionnaire. Fielding was completed in October 1998.

In 41% of the returned questionnaires, the form was completed by the superintendent or a person in the superintendent's office. In the remaining cases, the individual responsible for health education policy in the district (such as the curriculum director) usually completed the questionnaire. This occurred more commonly in larger districts, which are more likely to delegate administrative responsibilities.

A total of 825 school superintendents or their delegates responded, for an overall response rate of 68%—84% among large-enrollment districts (n=187), and 64% among small- and medium-enrollment districts (n=318 and n=320, respectively).

To adjust for nonresponse and for the enrollment size strata, we assigned a weight to responding school districts that inflated the number of cases to the actual number of eligible districts in the United States as a whole (13,560). We used the software package Stata to conduct tests of significance because the survey was based on a complex stratified sample. (Stata uses the unweighted number of cases and incorporates information from the sample weights and stratified design to inflate the standard errors for significance testing.)

To provide some context for the distribution of responses by school districts, we also examined some selected variables by the number of students in the United States. We created student weights by multiplying the number of students enrolled in each sampled district by the ratio of the number of students in the universe of all districts to the number of students in the responding sampled districts. This resulted in weighted estimates of all students in the universe of eligible school districts (43,276,146 students in districts that offer instruction in grade six or higher).

The questionnaire administered to district representatives defined "sexuality or abstinence education" as "any and all health education relating to human sexuality, including family life, abstinence until marriage, postponing sexual involvement, and avoidance of STDs or HIV and unintended pregnancy" (hereafter referred to as sexuality education).

While we defined policy as "any guidance that applies, district-wide, to sexuality education in the schools," some respondents crossed out policy and wrote in "practice." Some responded that they had no policy, and simply followed state directives; for our purposes, we considered these cases to have a policy.

We grouped the 825 districts according to the location categories defined in the sample-frame database, which classifies districts by their primary catchment area—the urban center of a metropolitan county (central city), the other areas of a metropolitan county (suburban) and areas completely outside metropolitan counties (nonmetropolitan). We also classified districts by four Census Bureau geographic regions—North, South, Midwest and West—and by nine subdivisions within these regions.

The level of missing data on overall sexuality education policy is quite low. For example, among the districts with a policy, only 4% did not supply details about how abstinence is taught. We did not impute missing data, but assumed that the responses on those few items that were missing would be similar to those of the responding districts. The item with the highest level of nonresponse was that asking districts with a sexuality education policy for the single most influential factor in establishing that policy (26%); in contrast, only 10% were unable to provide data on any factor influencing such policies.

We categorized districts along a continuum of how much emphasis their sexuality education program gives to abstinence promotion. We asked districts with a sexuality education policy which of the following best describes how their policy addresses abstinence:

•as one option in a broader educational program to prepare adolescents to become sexually healthy adults;

•as the preferred option for adolescents (when contraception is discussed, it is presented as an effective means of protecting against unintended pregnancy and STDs or HIV for sexually active individuals);

•as the only positive option outside of marriage (when contraception is discussed, its ineffectiveness in preventing pregnancy and STDs or HIV is highlighted); or

•as the only option outside of marriage (all discussion of contraception is prohibited). Since there were too few of these cases (36 unweighted districts, or 6% of all weighted districts with a sexuality education policy) to separately analyze them, we combined this category with the previous one to create a single abstinence-only category.

We also used this continuum to categorize districts into two general groups related to policies on instruction about contraception. Districts whose policy fell into the first two categories on the continuum were put into the "contraception as effective" category. In contrast, districts whose policy fell into the latter abstinence-only categories were grouped under "contraception as ineffective."

We also conducted multivariate logistic regression analyses to determine the combined impact of region, district size and metropolitan status on the likelihood that school districts would have a policy to teach sexuality education. Additionally, among districts having such a policy, we used multivariate analysis to examine the likelihood that their policy would be an "abstinence only" policy.

Results

Sexuality Education Policies

Among all U.S. school districts, more than two-thirds (69%) have a policy to teach sexuality education (Table 1). The remaining 31% leave policy decisions concerning sexuality education to individual schools within the district or to teachers.† A disproportionate number of students reside in districts that have policies to teach sexuality education. Among all U.S. students attending a district offering grade six or higher, 86% reside in school districts that have such a policy, while the remaining 14% attend schools in districts that leave these policy decisions to individual schools or to teachers (data not shown).

By region, school districts in the Northeast are the most likely to have a district-wide policy to teach sexuality education (86%, or 17 percentage points higher than the percentage for the country as a whole). Conversely, Midwestern school districts are the most likely to leave policy decisions to individual schools or teachers (41%). School districts in the South and West did not differ significantly from the U.S. average in the proportion having a policy to teach sexuality education.

These policies vary widely by subregions, however. For instance, while the South as a whole closely parallels the nation, almost all districts in the South Atlantic division have a policy (99%), while far fewer in the East South Central subdivision have one (40%).

Similarly, while the proportion of all districts in the Midwest having an explicit policy is significantly below the national average, this difference is true for the West North Central subdivision only (48% vs. 69%, p<.01), but not for the other Midwest subdivision (76% vs. 69%, a nonsignificant difference). The Northeast, meanwhile, is more homogenous in having a policy than the South or Midwest. School districts in both New England and the Middle Atlantic division are more likely than the nation as a whole to have district-wide policies (88% and 85%, respectively). Finally, the Mountain and Pacific subdivisions of the West do not differ significantly from the national average.

More than nine of 10 large-enrollment and medium-enrollment districts (91-95%) have a district-wide policy to teach sexuality education, compared with just 65% of small-enrollment districts. Because the vast majority of U.S. school districts have enrollments of fewer than 5,000 students, however, this percentage among the small-enrollment districts is not significantly different from the national average.

The proportion of school districts with a policy was significantly lower for those located in nonmetropolitan counties (60%) than for those in either central city (84%) or suburban counties (81%). The proportion of suburban school districts with a policy is significantly higher than the national average, while the proportion of nonmetropolitan districts is significantly lower than the nation as a whole. The proportion among central city districts does not differ from the national average.‡

Abstinence Policies

Among districts that have a policy to teach sexuality education (Table 2), 14% reported that their policy addresses abstinence as one option for adolescents to avoid pregnancy and STDs in a broader sexuality education program that includes discussion of contraception to prepare them to become sexually healthy adults (hereafter referred to as a comprehensive sexuality education policy). One-half of districts (51%) with a policy promote abstinence as the preferred option for adolescents; this policy allows contraception to be discussed as effective in protecting against unintended pregnancy and STDs or HIV (referred to as an abstinence-plus policy). Slightly more than one-third (35%) reported that their policy requires that abstinence be taught as the only option outside of marriage; discussion of contraception is either prohibited or its ineffectiveness in preventing pregnancy and STDs or HIV is highlighted (referred to as an abstinence-only policy). Thus, in terms of specific policy toward providing contraceptive information, 65% of districts with a policy allow discussions to portray contraception as effective in preventing pregnancy and STDs (the sum of the first two categories), while 35% either highlight contraception's ineffectiveness or prohibit discussion of contraception outright.

When all school districts in the country are taken into account—including those that do not have a policy to teach sexuality education—10% of U.S. school districts have a comprehensive sexuality education policy, 34% have an abstinence-plus policy, 23% an abstinence-only policy and 33% have no policy (data not shown). Among all U.S. students attending a district that includes grade six or higher, 9% are in districts that have a comprehensive sexuality education policy, 45% are in districts with an abstinence-plus policy, 32% in abstinence-only policy districts and 14% in districts that have no policy (data not shown).

There is considerable regional variation in how districts address the issue of abstinence. For instance, 55% of Southern school districts with a policy address abstinence as the only option for adolescents outside of marriage, a level 20 percentage points higher than the national average; in contrast, only 20% of districts in the Northeast with a policy have an abstintence-only policy—almost 15 percentage points below the national average. The South also has the lowest percentage (5%) of districts that direct that abstinence be taught as part of a comprehensive sexuality education program; while the percentage of Northeast districts that have a comprehensive policy is greater than the national average, this difference is not statistically significant. School districts in the Midwest and West do not differ significantly from all U.S. districts in how they address teaching abstinence.

In terms of regional subdivisions, the areas with the highest proportion having abstinence-only policies are both in the South—the West South Central subdivision (60%) and the South Atlantic subdivision (54%). (The third Southern subdivision, East South Central, does not differ from the national average.) While the proportion with an abstinence-only policy is lowest in New England (14%), the highest proportion of districts teaching abstinence as part of a comprehensive policy is found in the Middle Atlantic states (30%). New England districts are most likely to have a policy to teach an abstinence-plus curriculum (67%), although this proportion is not significantly different from the national average.

Districts' type of sexuality education policy does not vary appreciably by their enrollment size or metropolitan status, with the exception that large districts are significantly less likely than U.S. districts overall to treat abstinence as part of a comprehensive program.

Multivariate analyses indicate that when the effects of region and metropolitan status are taken into account, the relationship between district size and policy noted in Table 1 remains (first column of Table 3). Small-enrollment districts are about one-10th as likely as large ones to have such a policy (odds ratio, 0.11). Medium-sized districts are only about half as likely as large-enrollment districts to have a policy, but this difference is not statistically significant.

The relationship between region and district policy appears to be independent of the size or metropolitan status of a district. For example, net of other factors, school districts in the Midwest and in the South are significantly less likely than those in the Northeast to have a district-wide policy. The reduction in the likelihood of having a policy among districts in the West is not statistically significant, however.

Metropolitan status failed to independently affect the likelihood of having a policy, once the other variables were taken into account. This suggests that the finding in Table 1 that nonmetropolitan districts were less likely to have a policy was more a function of region and district size than of metropolitan status.

When we conducted among districts with a policy a multivariate analysis that controlled for region, enrollment size and metropolitan status, only region was independently associated with having an abstinence-only policy (Table 3). Southern districts were almost five times as likely as Northeastern districts to have a policy that teaches abstinence as the only option for unmarried adolescents. Midwestern districts were moderately more likely than Northeastern districts to have such a policy, but this association failed to reach statistical significance. After region was controlled for, district size and metropolitan status appeared to have no independent impact on whether a district has an abstinence-only policy.

Changes in District Policy

Among respondents who knew when their current policy was adopted (n=5,149 weighted districts), 53% said that their current policy was adopted after 1995, and another 31% said it was adopted between 1990 and 1995; only 16% said that their current policy predated 1990. There was no relationship between when a policy was adopted and the type of policy toward teaching abstinence (not shown).

Among respondents who knew whether their district's policy had replaced an existing one (n=5,920 weighted districts), almost one-quarter (23%) reported that their current policy had done so, while 77% indicated that their current policy had not replaced a prior policy.

Districts that indicated that their policy had been replaced were asked how their previous policy had addressed the teaching of abstinence following the same four-category scale. Among these districts, 52% said that their new sexuality education policy fell within the same general category as their former policy (the sum of the three descending diagonal cells in Table 4): The unchanged policy was to teach abstinence within a comprehensive program in 6%, as the preferred option for adolescents in 25% and as the only option in 21%. However, among districts that changed their policy, twice as many shifted toward a greater focus on abstinence as moved in the opposite direction. Thirty-three percent reported that their policy had changed from either a comprehensive to an abstinence-plus policy or from an abstinence-plus to an abstinence-only policy (the sum of the three cells above the diagonal), while just 16% reported that their policy had moved either from an abstinence-only policy to some other policy or from an abstinence-plus to a comprehensive policy (the sum of the three cells below the diagonal).

This shift among districts with replacement policies, however, had no net impact on the percentage of policies portraying contraception as effective or as ineffective. Fifteen percent of districts moved from having a policy in which contraception could be discussed positively (either a comprehensive or an abstinence- plus policy) to one in which contraception, if it is discussed at all, could only be portrayed negatively (an abstinence-only policy). Another 15% of districts, however, shifted from having an abstinence-only policy to a policy that permits contraception to be discussed as an effective means of preventing pregnancy and disease.

Finally, among the districts with replaced policies, there was virtually no net change in the total number of districts with abstinence-only policies (from 464 to 461, or a 0.6% decline). The major net shifts were in the number of districts that originally had had a comprehensive policy (from 301 to 189, a decline of 37%), and in the number of districts that had had an abstinence-plus policy (from 525 to 640, an increase of 22%).

Factors Influencing Policy

Respondents were asked to choose, from 11 possibilities,§ the single most important factor that influenced their district's current sexuality education policy (Table 5). One of just three factors (state directives, recommendations of special school board advisory committees or task forces, or school board actions) was named by at least three-quarters of districts, ranging from 78% of districts with abstinence-only policies to 88% of those with abstinence-plus policies. There were no significant differences in the percentage distributions according to the most influential factor between the three policy categories and the distribution for the nation as a whole.

On average, almost one-half of the districts (48%) cited state directives as the most influential factor. Special committees and school boards were cited as most influential about equally as often (18% and 17%, respectively). School boards were more likely than other factors to be considered as most important by districts with an abstinence-only policy, but this proportion did not differ significantly from that among all U.S. districts.

Respondents were also asked to indicate from the same list whether any of the factors had influenced their current policy. Districts reported an average of 2.6 factors. In general, the responses followed the same pattern as that created by the most influential factor, with state directives being the most frequently cited influential factor of all possibilities (74%). However, districts cited school board actions more frequently than they did special committees (63% vs. 36%). Predictably, the proportions checking community and teacher support for abstinence as influential factors were higher in districts with abstinence-only policies, whereas community support for broader sexuality education was more prevalent in districts that have comprehensive policies.

Finally, respondents were asked how supportive they thought the community at large was of their district's current policy on sexuality education. The most common response (53%) was that the community was "generally silent" on the issue; 41% reported that their community "strongly supports" the current policy, 5% that the community is divided and fewer than 1% that it is "generally opposed." The level of community support within each of the three sexuality education policy subgroups did not differ significantly from the nation as a whole. However, districts in which abstinence is presented as the only option outside of marriage for adolescents were somewhat more likely than other districts to have higher levels of community support for their policy (at least according to the school superintendent), and communities in these abstinence-only districts were less likely to be "generally silent" on the issue.

Discussion

By 1998, more than two out of three public school districts in the United States had adopted a district-wide policy to teach sexuality education. Most of those policies were developed in the 1990s, during a period of intense debate, not only in many local communities but also in state capitals and Congress, about the relative merits of abstinence promotion—and, specifically, abstinence-only promotion—versus more "comprehensive" approaches to sexuality education for young people. Yet the impact of those debates at the local level and the trend in the national debate are not especially easy to assess.

On the one hand, the overwhelming majority of policies now require that abstinence from sexual intercourse be promoted—either as the preferred option for adolescents or as the only option outside of marriage. Few districts stipulate that abstinence is to be presented as one option in a broader educational program to prepare adolescents to become sexually healthy adults.

On the other hand, among districts that adopted new policies, the newer policies do not appear to be more "conservative" regarding how abstinence is treated than those adopted earlier in the decade. Moreover, the bulk of the movement among those districts that switched from one policy category to another appears to have been away from comprehensive sexuality education policies toward abstinence-plus policies; indeed, there was no net movement toward the most extreme abstinence-only policies.

Because abstinence-plus policies allow contraception to be discussed as an effective means of providing protection against pregnancy and disease, adding these districts to those presenting abstinence as part of a comprehensive educational program for adolescents indicates that two-thirds of all districts that have a policy permit positive discussions of contraception, notwithstanding the extent to which those policies also promote abstinence.

Still, more than one-third of districts with a policy to teach sexuality education require that abstinence be taught as the only option outside of marriage; under the vast majority of these policies, contraception may only be discussed in a way that highlights its shortcomings. (A small percentage of these districts prohibit discussion of contraception outright.) Despite considerable regional variation, there is no region of the country in which the proportion of districts with abstinence-only policies is negligible—one-fifth of districts in the Northeast with a policy, more than one-quarter in the West, more than one-third in the Midwest and more than one-half in the South.

The exclusive focus on abstinence promotion in these policies is troubling, in light of the dearth of research demonstrating that the abstinence-only approach is effective in delaying young people's sexual initiation. This lack of documentation stands in sharp contrast to the growing weight of evidence showing that broader educational approaches appear to delay sexual initiation.¹² Moreover, while more comprehensive approaches also have been shown to encourage greater use of contraceptives when young people eventually begin to have intercourse, the impact of abstinence-only programs on youth's subsequent contraceptive use has yet to be addressed.¹³ By emphasizing the failure rates of contraceptive methods or by permitting no discussion about contraception at all, abstinence-only efforts might discourage effective contraceptive use and thereby put individuals at greater risk of unintended pregnancy when they become sexually active.

With more than 70% of young Americans aged 18-19 having initiated sexual intercourse,¹⁴ the provision of adequate and accurate information about contraception—even while supporting the choice of young people who are delaying sexual initiation—should continue to be a high national priority. This is especially important if national declines in teenage pregnancy rates experienced from the late 1980s through the mid-1990s are to be sustained. A recent analysis indicates that those declines were associated not only with a modest increase in the proportion of young people who had never had sexual intercourse, but even more so with a lower likelihood of becoming pregnant among sexually experienced teenagers.¹⁵

In that regard, recent legislative efforts by California and Missouri to ensure that information presented to students in the context of sexuality education is "medically accurate," even within a framework that presents abstinence as the preferred choice, are encouraging. The Missouri legislation was supported by an alliance of organizations from opposite ends of the political spectrum that share the goals of reducing rates of teenage pregnancy and of STDs.¹⁶

Finally, it should be emphasized that this study was initiated in early 1998, well before states began implementing any abstinence-only promotion efforts of their own following enactment of the federal welfare reform legislation that guaranteed federal funds for school- and community-based programs over a five-year period. Additional research clearly is warranted to monitor and evaluate the extent to which that law and its implementation may be providing the impetus for additional changes in school district policy on the teaching of abstinence in the context of sexuality education.