Table 3.

Respondent driven sampling (RDS) methodology features, description of those features, and considerations for applying those features when conducting surveys of adolescent key populations (AKP).

Feature	Current implementation of features	Considerations of features when sampling AKPa
Pre-survey research	Conduct pre-survey research to assess populations’ social networks, RDSb acceptability, and logistics.	Organizations working with AKP can help identify sites where AKP spend time, meet friends, find new sex partners, or buy or sell drugs. Sites may or may not be the same as those frequented by adult key populations.
Seeds	Seeds are the initial participants that start the recruitment from within the network of interest.	In surveys of adolescents and adults together, select an ample number of AKP seeds who are more likely to recruit other AKP. In surveys of AKP only, special considerations may be needed to diversify seeds. Some adult diversifications, such as marital status or occupation, may not be applicable to AKP. Consider other diversity factors such as whether they live at home or not, are sexually active or not, in school or not, able to read and write or not, etc.
Incentives	A nominal incentive, usually in the form of local currency, is usually provided to participants who complete the survey process and for recruiting their peers.	Make sure a country’s research ethical standards allow incentives to be provided to adolescents. Incentives used for adults may not be appropriate for adolescents. Involve organizations working with AKP to select appropriate incentives. Ensure that incentives do not induce participation (eg, if it were not for the incentive provided, the participant would not enroll in a survey or would not withdraw from the survey early, given his or her better judgment).
Measuring social network size	For RDS analysis, the number of eligible people that each subject “knows” and has seen during a specified period of time (eg, 2 weeks) is needed. This question is open-ended and gaining accurate responses can be difficult, especially from someone uncomfortable with counting.	Use special probing techniques to assist with and encourage accurate reporting of social network sizes. Some adolescents may need help with coming up with a number.
Interview sites	RDS usually requires recruits to be present at an interview site to undergo the survey.	Interview sites should be easy to access, safe, and comfortable for AKP. In surveys of adolescents and adults together, consider special hours for AKP to enroll when adult participants are not present. Avoid interview sites located close to schools, homes, police stations, prisons, sex work solicitation areas, high crime areas, etc. Consider what to do with family members who bring AKP to participate in a survey. Consider the use of WebRDS [45-47], whereby peers recruit their peers via messaging and email.
Staffing	RDS surveys usually have numerous staff members, including someone to screen for eligibility, interviewers, pre- and post-test HIVc counselors, coupon managers, and so on.	Ensure staff members are trained to interact with adolescents and can recognize signs of distress and able to respond appropriately. Retain a social worker familiar with AKP at the survey site to respond to difficult situations. Ensure staff members understand the consent form and are trained to acquire informed consent from adolescents and/or their parent or guardian without coercing the adolescent. Have a staff member to engage adolescents with activities while they are waiting for an interview. Support staff who may be emotionally impacted by their exposure to adolescents who sell sex, use drugs, have experienced violence, are living with HIV, or have other risks and vulnerabilities.
Coupons	Coupons are used by participants to recruit peers. Coupons convey important information about the survey location, operation hours, etc.	Use simple coupon designs and words that are easily understandable to adolescents. Use pictures in place of words or other creative means to convey important coupon information. If using WebRDS [45-47], coupons can be sent through Web-based applications.

^aAKP: adolescent key populations.

^bRDS: respondent driven sampling.

^cHIV: human immunodeficiency virus.