Respondent-Driven Sampling Study of Commercial Sexual Exploitation of Children (CSEC) in Kampala, Uganda

Respondent-Driven Sampling Study of Commercial Sexual Exploitation of Children (CSEC) in Kampala, Uganda

Respondent-Driven Sampling Study of Commercial Sexual Exploitation of Children (CSEC) in Kampala, Uganda

Executive Summary

This study is the second of a two-phase study on the commercial sexual exploitation of children (CSEC) in Kampala, Uganda. The first phase of this study took place during March through April 2021, and the second phase took place during July through September 2022. The study uses a repeated cross-sectional design; the samples at Time 1 and Time 2 are independent and were selected using the same methodology.

The objectives of these studies were to (1) create population-based measures of CSEC to explore the change in CSEC prevalence over time and (2) understand the working conditions of the children involved in CSEC. The study was implemented by ICF and Makerere University’s Department of Social Work and Social Administration with funding from the Global Fund to End Modern Slavery through the Program to End Modern Slavery in the U.S. Department of State’s Office to Monitor and Combat Trafficking in Persons.


The prevalence of CSEC and a thorough identification of vulnerabilities and risk factors have not been fully studied in Kampala. No prior studies, other than the first phase of the present study, offer an estimate of the prevalence of CSEC in Kampala. A capture-recapture study estimates the number of females age 15 or older who are engaged in commercial sex in Kampala to be about 9,000; however, this study did not include individuals younger than 15 and does not estimate the proportion who are ages 15 to 17.1 There is no available estimate of the number of males who are engaged in commercial sex.


The samples at both time points were recruited using respondent-driven sampling (RDS), a network based sampling method that overcomes the traditional biases associated with similar approaches (e.g., chain-referral and snowball sampling) by approximating probability sampling methods and allowing for the calculation of selection probabilities and survey weights. RDS weights reflect the varying sizes of respondents’ networks as established in RDS theory, which adjusts for recruitment biases. Initial participants in an RDS study (i.e., seeds) are recruited through convenience sampling methods. Each of these seeds recruits peers by referral, allowing researchers to access members of typically hard-to-reach populations who may not otherwise be accessible.2

The initial respondents (seeds) in this study were recruited with support from four local organizations that work with and provide support to survivors of commercial sexual exploitation (CSE) in Kampala. At Time 2, there were 12 seeds ages 15 through 17 and 12 seeds age 18 and older. The majority of the seeds (17) are female and 7 are male. The seeds recruited additional respondents, who then recruited other respondents. Respondents were offered a maximum of three coupons with which to refer other respondents. To encourage participation and the referral of peers, respondents were offered an incentive for the completion of an interview and for referring other respondents who successfully completed an interview. The final sample at Time 2 includes 240 respondents age 15 or older who live or work in Kampala and engaged in CSE in the past year.


A general limitation of RDS methods is that while weighting compensates for the reduced probability of capturing eligible individuals who are not well connected, the approach cannot cover persons who are not connected at all.

Due to logistical constraints, this study had a relatively large number of seeds and, therefore, relatively short referral chains. Half of our seeds were minors (under age 18); the predominance of minors as seeds may skew the estimated prevalence of minors among individuals engaged in the sex industry generated using our weighted sample. More than one-fourth (29%) of our seeds are male, which may skew the estimated prevalence of males among children engaged in CSEC.

Weights and estimates based on RDS are premised on a semi-probability sampling method (at best). Therefore, it is difficult to compute the variance of the RDS sample estimates, including the estimated prevalence. Estimated standard errors involve approximations related to the RDS assumptions.

These limitations and the different composition of the seeds at Time 1 and Time 2 make it challenging to determine the comparability of the estimates generated at each time point. Without reliable benchmark population estimates, our qualified conclusion that the changes in estimates reflect a true change in the population should be interpreted as suggestive rather than certain.


The findings from the Time 2 (2022) study are summarized as follows:

Prevalence: We estimate that approximately 27.6 percent of individuals engaged in the sex industry in Kampala are under age 18. We find no change in the prevalence of children in CSE between Time 1 (2021) and Time 2 (2022).

Gender: At Time 2, the estimated percentage of children engaged in the sex industry in Kampala who are male was 37 percent compared with 23 percent at Time 1. The difference between the estimates is not statistically significant and, therefore, we find no change in the proportion of males among children in CSE between Time 1 and Time 2.

Means of connecting with clients: Just over one-fourth of people engaged in the sex industry (27%) knew the client before the first commercial sexual activity. The percentage is more than twice as high for children (37%) compared with adults (16%). Regarding the most recent commercial sexual activity, children were again nearly twice as likely to know the client beforehand compared with adults (32% versus 18%, respectively). There also were notable differences by age in how the client was identified. Nearly 70 percent of adults met the client in a bar or on the street compared with 42 percent of children. A greater proportion of children compared with adults reported having an existing friendship with the client or being neighbors with the client.

Decision making and coercion: A significant proportion of those engaged in the sex industry—nearly half of children (42%) and one-third of adults (34%)—worked for a pimp or broker at least sometimes. Despite this, individuals engaged in the sex industry appear to have a fairly high level of autonomy. Nearly all respondents are a decision maker regarding whether they do sexual things and with whom they do sexual things. The decision about where the respondent will go for the sexual activity is less under the control of the person engaged in the sex industry, particularly for children. Only around half of children reported that they are a decision maker regarding where they will go (compared with 85% of adults). For all of these decisions, a greater proportion of children than adults reported that a pimp or broker was a decision maker.

Close to one-fourth of individuals involved in CSE are pressured or forced to do sexual things. Children reported feeling pressured or forced at almost twice the rate reported by adults (32% for children versus 19% for adults). Nearly one-fifth (19%) of all individuals involved in CSE have felt that they would be hurt if they did not do something they were told to do. One-tenth (11%) of all individuals involved in CSE have been hurt because they did not do something they were told to do.

Possible effects of COVID-19 restrictions on the Kampala sex industry: Our findings show an increase in new entrants to the sex industry during the pandemic restrictions. A larger number of people may have been driven into the industry during the pandemic restrictions because they had no other means of support during the closures.

At Time 1, 14 percent always worked for a pimp or broker; however, at Time 2, only 2 percent always worked for a pimp or broker. It is possible that individuals who operated exclusively through a pimp or broker during the pandemic restrictions may have begun operating independently since the restrictions have been lifted. In addition, there was a large decrease in the percentage of respondents who sometimes or always feel pressured or forced to do sexual things, from 45 percent at Time 1 to 23 percent at Time 2. This decrease can be partially attributed to the decreased involvement with pimps and brokers; however, it also likely relates to the decreased vulnerability of individuals once the pandemic restrictions eased and other avenues of self-support re-emerged.

Impact of ending the COVID-19 restrictions: Respondents agreed that the easing of restrictions improved their ability to support themselves through commercial sex. Respondents indicated that the reopening of bars and similar establishments has made it easier to find clients, and the reopening of public transport has made it easier to get to clients. Many respondents noted that the decreased police presence and the end of the curfew makes their work, which takes place primarily at night, much easier and safer.

Recommendations by child respondents: More than three-fourths (78%) of children engaged in the sex industry would like to be provided with employment support. Nearly half of children (44%) mentioned educational support. One-third (34%) of children discussed cash transfers. A small number of children stated the need for health support and the prosecuting of traffickers and rapists.

Read or download full report here.