Mar 14, 2016


 This is an online sexual education program implemented in public junior high schools throughout Colombia.  The program occurs during the school year.  The program focuses on condom use and knowledge, sexually transmitted infection (STI) prevention, violence prevention, and pregnancy prevention, as well as attitudes about safe sex. An experimental evaluation found impacts on knowledge and attitudes.


 Target population: Colombian high school students

This is a computer-based sexual education curriculum. The largest Colombian NGO that works on issues of sexual and reproductive health (and, for a time, the only importer of condoms in the country) designed the program.  Through online modules, the curriculum aims to reach a high number of high school students rapidly to teach about STI symptoms and causes, awareness of sexual violence, STI prevention, pregnancy prevention, and proper condom use.  It also aims to change behavior and increase condom use by providing vouchers for free condoms, redeemable at a local health center. The program can be accessed anywhere, at any time, and online counselors are also available to answer questions.  There are five modules that require a total of 11 weeks to complete; students in each of the schools implementing the program are given 1.5 hours each week during school hours to complete the program.  However, because the program is online, students may complete them at any time, using a personal log-in.  The program leaders estimate that the program costs approximately $1,000 for 68 students to receive the program.


 Chong, A., Gonzalez-Navarro, M., Karlan, D., & Valdivia, M. (2015). Effectiveness and spillovers of online sex education: Evidence from a randomized evaluation in Colombian public schools.” NBER Working Paper.

 Evaluated population: One hundred thirty-eight classrooms in 69 schools, with a total of 4,599 students, were included in this evaluation.  The majority of schools started their school year in January, were a single-shift school (did not have students attend in scheduled waves), had a morning shift, and taught sex education.  Additionally, most were in smaller cities (fewer than 600,000 residents), had 3 9th-grade classrooms, had an average of 37 or 38 students in each class, and had 36 to 38 computers per school.  Approximately 50-60 percent of the participants were female, most were not sexually active, were aged 15 on average, had parents with approximately 12-13 years of schooling, did not have a computer at home but did own a cellphone, used the internet, and felt that religion was important.

Approach: Randomization took place at three levels in this evaluation.  First, schools were randomized to treatment or control groups.  Twenty-three schools were assigned to be control schools, with two classrooms in each school participating, while 46 were assigned to be treatment schools.  Within the treatment schools, one classroom was randomly assigned to treatment, and one to “spillover” status. The spillover classroom did not receive the training, but was included in order to test whether having peers who learn about sex education might spill over and impact students who did not receive the curriculum.  This meant that there were 46 classrooms control classrooms, 46 intervention classrooms, and 46 spillover classrooms.  Knowledge outcomes in the five domains described above were measured, as was condom use.  In order to create a condom-use measure that was more behavioral than self-reported, participants were given a voucher to get six free condoms (a value of approximately $5) at a local clinic.  In order to assess whether transportation access impacted the use of these vouchers, half of all treatment participants were also randomly offered reimbursement for bus fare.

At baseline, all three groups were statistically equivalent, except with regard to gender.  There were more boys in the control group than in the intervention or spillover groups (49 percent, versus 41 percent and 40 percent, respectively).

Data were collected before the program (at the beginning of the school year), immediately following the program, and six months later (at the end of the school year).

Results: Aggregate scores for knowledge (all of the scales combined) increased in the intervention group compared to the control group (measured one week after the completion of the program), and the increase was maintained after six months. Each individual scale – knowledge of symptoms and causes of STIs, sexual violence, STI prevention, pregnancy prevention, and condom use – also showed improvements in the intervention group compared to the controls.  There was no clear evidence of spillover.  This meant that in the 46 classrooms that did not receive the treatment, but where another classroom in the same school did receive it, there was no evidence of impact on the non-treated students.

Attitudes also improved in the intervention group.  Improvements were found for attitudes related to condom use, the need to report sexual violence, and attitudes that the authors defined as conservative (beliefs about avoiding multiple sexual partners, having a partner of a similar age, and waiting for emotional maturity to engage in sexual activity).



 Chong, A., Gonzalez-Navarro, M., Karlan, D., & Valdivia, M. (2015). Effectiveness and spillovers of online sex education: Evidence from a randomized evaluation in Colombian public schools.” NBER Working Paper.

KEYWORDS: high school, adolescent, condom use, STD/HIV/AIDS, Hispanic/Latino, Sexual Activity,  Teen Pregnancy

Program information last updated on 3/14/2016