This study used a correlational approach to assess whether abstinence-only education is effective in reducing U.S. teen pregnancy rates. Correlation can be due to causation, but it can also be due to other underlying factors, which need to be examined. Several factors besides abstinence education are correlated with teen pregnancy rates. In agreement with previous studies, our analysis showed that adjusted median household income and proportion of white teens in the teen population both had a significant influence on teen pregnancy rates. Richer states tend to have a higher proportion of white teens in their teen populations, tend to emphasize abstinence less, and tend to have lower teen pregnancy and birth rates than poorer states. A recent study 
found that higher teen birth rates in poorer states were also correlated with a higher degree of religiosity (and a lower abortion rate) at the state level. Medicaid waivers have previously been shown to reduce teen pregnancy rates 
, but our analysis shows that they do not explain our main result, the positive correlation between abstinence education level and teen pregnancy rates.
After accounting for other factors, the national data show that the incidence of teenage pregnancies and births remain positively correlated with the degree of abstinence education across states: The more strongly abstinence is emphasized in state laws and policies, the higher the average teenage pregnancy and birth rate. States that taught comprehensive sex and/or HIV education and covered abstinence along with contraception and condom use (level 1 sex education; also referred to as “abstinence-plus” 
, tended to have the lowest teen pregnancy rates, while states with abstinence-only sex education laws that stress abstinence until marriage (level 3) were significantly less successful in preventing teen pregnancies. Level 0 states present an interesting sample with a wide range of education policies and variable teen pregnancy and birth data 
. For example, several of the level 0 states (as of 2007) did not mandate sex education, but required HIV education only (e.g. CT, WV) 
. Only three of the level 0 states (IA, NH and NV) mandated both sex education and HIV education, but one of them (NV) did not require that teens learn about condoms and contraception. This state (NV) has the highest teen pregnancy and birth rates in that group (). Nevada is also one of only five states (with MD in level 0, CO in level 2, and AZ and UT in level 3) that required parental consent for sex education in public schools instead of an opt-out requirement that is present in all the other states 
The effectiveness of Level 1 (comprehensive) sex education in our nation-wide analysis is supported by Kirby's meta-analysis of individual sex education programs 
, Underwood et al. 's analysis of HIV prevention programs 
, and a recent review by the CDC taskforce on community preventive services 
. All these studies suggest that comprehensive sex or HIV education that includes the discussion of abstinence as a recommended behavior, and also discusses contraception and protection methods, works best in reducing teen pregnancy and sexually transmitted diseases.
Individual research studies
Despite large differences between individual research studies that evaluate specific sex education programs (e.g. sample size, approaches to sex education studied, selection of participants, choice of control groups, types of data, control for cross-talk between students outside of class, etc.), several case studies show that abstinence-only education rarely has a positive effect on teen sexual behavior 
. One of the few exceptions is the recent study by Jemmott et al. 
on black middle school students in low-income urban schools: after receiving 8 hours of abstinence education as 12 year olds, significantly more students (64/95) reported to be abstinent after 24 months when compared to (control) students who received 8 hours of health education (without any form of sex education: 47/88; Fishers exact test, p
0.037), or students who received 8 hours of safe-sex education (without an abstinence component: 41/85, Fishers exact test, p
0.007). However, there was no significant difference in abstinence behavior between students who had received abstinence education (64/95) and students who received 8 hours of comprehensive sex education (combining sex education with abstinence education: 57/97; Fishers exact test, p
0.138). These two groups also did not differ in rates of reported unprotected sex (8/122 versus 8/115) or use of condoms (25/33 versus 29/37) in the previous 3 months. The abstinence-only intervention in that study was unique in that it increased knowledge about HIV/STD, emphasized the delay of sexual activity, but not necessarily until marriage, did not put sex into a negative light or use a moralistic tone, included no inaccurate information, corrected incorrect views, and did not disparage the use of condoms 
. As a result, as pointed out by the authors, this successful version of abstinence education would not have met the criteria for federal abstinence-only funding 
. While promoting an alternative and more effective form of abstinence education, these results also support Kirby's findings 
and the data in the present study that comprehensive sex education that includes an abstinence (delay) component (level 1), is the most effective form of sex education, especially when using teen pregnancy rates as a measurable outcome.
Individual research studies also show that teaching about contraception is generally not associated with increased risk of adolescent sexual activity or sexually transmitted diseases (STDs) 
as suggested by abstinence-only advocates, and adolescents who received comprehensive sex or HIV education had a lower risk of pregnancy and HIV/STD infection than adolescents who received strict abstinence-only or no sex education at all in the U.S. and in other high-income countries 
Abstinence-only education: public opinion and associated costs
Despite the data showing that abstinence-only education is ineffective, it may be argued that the prescribed form of sex education represents the underlying social values of families and communities in each state, and changing to a more comprehensive sex education curriculum will meet with strong opposition. However, there is strong public support for comprehensive sex education 
. Approximately 82% of a randomly selected nationally representative sample of U.S. adults aged 18 to 83 years (N
1096) supported comprehensive programs that teach students about both abstinence and other methods of preventing pregnancy and sexually transmitted diseases. In contrast, abstinence-only education programs, received the lowest levels of support (36%) and the highest level of opposition (about 50%).
In addition to the federal and state funds spent on abstinence-only (level 3) education, there are other costs associated with the outcomes of failed sex education and family planning. When deciding state policies on sex education, State legislators should consider these additional costs. For example, based on estimates by the National Campaign To Prevent Teen and Unplanned Pregnancy 
, teen child bearing (compared to first birth at 20 years or older) in the U.S. cost taxpayers (in direct and indirect costs) more than $9.1 billion in 2004.
Our data show that education (% of high school graduates taking the SAT) was not correlated with teen pregnancy rates, but it was positively correlated with teen abortion rates and negatively correlated with teen birth rates. These data can be interpreted in two ways: (1) pregnant teens who give birth are less likely to finish high school and go on to college (i.e. pregnancy affects education). This is supported by a recent report 
that showed that teen mothers are more likely to drop out of school: 51% of teen mothers earned their high school diploma by age 22, compared to 89% of women who had not given birth as teens. (2) teens who are motivated to go to college are not necessarily less likely to get pregnant, but more likely to abort their pregnancies (i.e. educational goal affects the decision of whether to carry a pregnancy to term).
As pointed out by the Society for Adolescent Medicine, the abstinence-only approach (as stressed by level 3 state laws and policies and funded by the federal abstinence-only programs) is characterized by the withholding of information and is ethically flawed 
. Abstinence-only programs tend to promote abstinence behavior through emotion, such as romantic notions of marriage, moralizing, fear of STDs, and by spreading scientifically incorrect information 
. For example a Congressional committee report found evidence of major errors and distortions of public health information in common abstinence-only curricula 
. As a result, these programs may actually be promoting irresponsible, high-risk teenage behavior by keeping teens uneducated with regard to reproductive knowledge and sound decision-making instead of giving them the tools to make educated decisions regarding their reproductive health 
. The effect of presenting inadequate or incorrect information to teenagers regarding sex and pregnancy and STD protection is long-lasting as uneducated teens grow into uneducated adults: almost half of all pregnancies in the U.S. were unplanned in 2001 
. Of these three million unplanned pregnancies, ~1.4 million resulted in live births, ~1.3 million ended in abortion, and over 400,000 ended in a miscarriage 
at a financial cost (direct medical costs only) of ~$5 billion in 2002 
The U.S. teen pregnancy rate is substantially higher than seen in other developed countries () despite similar cultural and socioeconomic patterns in teen pregnancy rates 
. The difference is not due to the onset of sexual activity 
. Instead, the main factor seems to be sex education, especially with regard to contraception and prevention of STDs 
. Sex education in Europe is based on the WHO definition of sexuality as a lifelong process, aiming to create self-determined and responsible attitudes and behavior with regard to sexuality, contraception, relationships and life strategies and planning 
. In general, there is greater and easier access to sexual health information and services for all people (including teens) in Europe, which is facilitated by a societal openness and comfort in dealing with sexuality 
, by pragmatic governmental policies 
and less influence by special interest groups.
While states with comprehensive sex education have lower teen pregnancy rates, even in these states rates are much higher than seen in Europe 
. This is likely influenced by the fact that U.S. state laws and policies generally do not require that sex and STD education is taught in all schools, but only provide guidelines if local school boards decide to teach it 
. For example, as of August 1, 2011, only 20 states mandated sex education, and 32 states mandated HIV education in their schools 
. In addition, even states with comprehensive sex education laws or policies (level 1) received federal funding for individual abstinence-only education programs in 2005: total federal funds 
averaged ~$14 per teen in level 1 states compared to ~$21 per teen in level 2 and 3 states 
. An important first step towards lowering the high teen pregnancy rates would be states requiring that comprehensive sex education (with abstinence as a desired behavior) is taught in all public schools. Another important step would involve specialized teacher training. Presently the sex education and STD/HIV curricula are often taught by faculty with little training in this area 
. As a further modification, “sex education” could be split into a coordinated social studies component (ethics, behavior and decision-making, including planning for the future) and a science component (human reproductive biology and biology of STDs, including pregnancy and STD prevention), each taught by trained teachers in their respective field.
As parents, educators or policy makers it should be our goals that (1) teens can make educated reproductive and sexual health decisions, that (2) teen pregnancy and STD rates are reduced to the rates of other developed nations, and that (3) these trends are maintained through the teenage years into adulthood. One possibility for achieving these goals is a close alignment and integration of sex education with the National Science Standards for U.S. middle and high schools 
. In addition, the Precaution Adoption Process Model
() advocated by the National Institutes of Health 
offers a good basis for communication and discussions between scientists, educators, and sex education researchers, and could serve as a reference for measuring progress in sex education (in alignment with the new evidence-based Teen Pregnancy Prevention Initiative). In addition, it could be used as a communication tool between sex education teachers and their students. It should be our specific goal to move American teens from Stages 1 or 2 (unaware or unengaged in the issues of pregnancy and STD prevention) to Stages 3–7 (informed decision-making) by providing them with knowledge, understanding, and sound decision-making skills (). For example, a recent study 
attributes 52% of all unintended pregnancies (teenagers and adults) in the U.S. to non-use of contraception, 43% to inconsistent or incorrect use, and only 5% to method failure.
The Precaution-Adoption-Process Model.
Our analysis adds to the overwhelming evidence indicating that abstinence-only education does not reduce teen pregnancy rates. Advocates for continued abstinence-only education need to ask themselves: If teens don't learn about human reproduction, including safe sexual health practices to prevent unintended pregnancies and STDs, and how to plan their reproductive adult life in school, then when should they learn it, and from whom?