Recruiting Adolescent Research Participants: In-Person Compared to Social Media Approaches

Abstract

Recruiting adolescent participants for research is challenging. The purpose of this study was to compare traditional in-person recruitment methods to social media recruitment. We recruited adolescents aged 14–18 years for a pilot physical activity intervention study, including a wearable physical activity tracking device and a Facebook group. Participants were recruited (a) in person from a local high school and an adolescent medicine clinic and (b) through social media, including Facebook targeted ads, sponsored tweets on Twitter, and a blog post. Data collected included total exposure (i.e., reach), engagement (i.e., interaction), and effectiveness. Effectiveness included screening and enrollment for each recruitment method, as well as time and resources spent on each recruitment method. In-person recruitment reached a total of 297 potential participants of which 37 enrolled in the study. Social media recruitment reached a total of 34,272 potential participants of which 8 enrolled in the study. Social media recruitment methods utilized an average of 1.6 hours of staff time and cost an average of $40.99 per participant enrolled, while in-person recruitment methods utilized an average of 0.75 hours of staff time and cost an average of $19.09 per participant enrolled. Social media recruitment reached more potential participants, but the cost per participant enrolled was higher compared to traditional methods. Studies need to consider benefits and downsides of traditional and social media recruitment methods based on study goals and population.

Introduction

Adolescents, particularly those who engage in risky health behaviors, remain a challenging population to recruit for research studies.^1–3 Traditional in-person recruitment methods for adolescents often involve visiting schools or engaging patients in clinics. These traditional recruitment methods involve costs of transportation for research staff to the sites and staff time, which consists of time spent building relationships to gain access to recruitment sites, as well as conducting recruitment by engaging with potential participants at these sites.

Adolescents' connection to social media has brought new opportunities to researchers to utilize social media tools to reach and recruit adolescents for research studies. Social media use among adolescents is nearly ubiquitous, with over 97% reporting use of social media and approximately a quarter of teens describing themselves as “constantly connected.”⁴ To date, the majority of studies that describe Facebook recruitment efforts and outcomes have focused on adults. These studies often describe use of Facebook advertisements to recruit targeted populations such as adult smokers,⁵ pregnant women, and men who have sex with men.^6,7

Costs involved in social media recruitment vary widely; one study of adults reported cost per participant recruited as less than $2 for an online survey study.⁸ However, a nutrition education study reported cost per adult participant recruited as nearly $20.⁹ Thus, it remains unclear whether social media advertisements offer an advantage in reach, engagement, and enrollment compared to in-person methods, as well as differences in cost. The purpose of this study was to describe the use of social media to recruit adolescents for an intervention study compared to traditional in-person recruitment methods at a school and clinic.

Methods

This study was conducted in the setting of a pilot randomized controlled trial. Data were collected between May and September 2014. The Western Institutional Review Board approved this study.

Study design

This pilot randomized controlled trial was designed to test feasibility, acceptability, and use of an intervention for increasing physical activity. Participants that enrolled were randomized to intervention or control groups. The participants in the intervention group were given a mobile physical activity tracker (Fitbit) to use for 4 weeks. The Fitbit Flex is a wrist-worn tracker that provides estimates of steps, energy expended, and distance traveled. Fitbit data and users' personalized goals can be provided in real time throughout the day and visualized on users' smart phones, tablets, and computers. Intervention participants were also invited to join a Facebook group.¹⁰

Recruitment

The goal of our study was to recruit high-school age adolescents between the ages of 14 and 18 years. While the developmental stage of adolescence encompasses a broad age range between 10 and 19 years,¹¹ we focused on high school teens for several reasons. First, we wanted to recruit teens who we hypothesized could utilize the Fitbit independently. Second, we wanted to recruit teens who met the minimum age requirement of 13 years to use Facebook.

In-person recruitment methods included recruitment from a local high school and an adolescent medicine clinic. For the high school, recruitment efforts began with contacting the high school to identify a staff member who would engage in communication about recruitment. This involved e-mails, phone calls, and a site visit to obtain approval to recruitment from the school district. A memorandum of understanding was drafted to outline the recruitment processes and signed by the school district and the lead researcher. On the day in which recruitment was scheduled, two project staff drove to the school and set up a recruitment table and a sign during the three school lunches. Research staff handed out flyers describing the study and engaged youth to discuss the study and answer questions. Interested students could sign up and later receive a phone call to complete parental consent.

Clinic recruitment involved obtaining approval from the head of the clinic; this involved several e-mails. On the 2 days that were selected as recruitment days, two research staffs were present in the clinic waiting room for the morning and afternoon clinic sessions. Staff again set up a recruitment table and sign and engaged in similar processes of handing out flyers and engaging youth and parents to discuss the study and answer questions.

Social media recruitment involved three approaches. The first approach was the design and posting of a Facebook advertisement (ad). We designed the ad to be appealing to adolescents and similar to other teen-targeted ads we identified on Facebook. We consulted with study staff with background in marketing and communication for its design and sought input from adolescents in reviewing draft ads. This process also involved several interactions with the Institutional Review Board (IRB) to obtain approval. The IRB initially refused to approve an ad that included mention of an incentive. Thus, Facebook ad 1 did not include mention of a study incentive. This ad was posted for one week and performed poorly. After additional interaction with the IRB in which we described the poor performance of the ad, Facebook ad 2 was approved by the IRB and posted for one week. Each ad was posted on Facebook for one week as a “click to pay” ad; thus, the account was not charged unless a viewer clicked the ad. The ad was specified to be displayed to adolescents aged 14–18 years in King County, WA.

The second approach attempted was a Twitter ad, which we designed to be identical to Facebook ad 2, targeting the same group and posted for one week. The final approach was through a blog post. The “Teenology101” blog is written by a Seattle Children's adolescent medicine physician and targets parents of teens. For this blog, we worked with the blog author to guest-author a post about physical activity and include a recruitment message about the study at the end of the blog.

All potential participants were directed to fill out an online screening survey. Telephone followup was done to obtain consent and enroll participants.

Measures

Measures for this study were designed to assess exposure, engagement, and enrollment for each recruitment method, as well as differences in cost.

Exposure

To measure reach for in-person approaches, we tracked how many potential participants were present at the school lunches and in clinic on the days in which recruitment took place. To measure social media reach, we assessed the number of ad views on Facebook and Twitter as reported by those companies, as well as the number of viewers of the blog post in the 7 days after it was posted as measured by Google Analytics.¹²

Engagement

To measure engagement for in-person approaches, we recorded how many people at recruitment sites took a study recruitment flyer. To measure engagement in social media, we used data analytic tools on each social media site. For Facebook ads, reported “click throughs”¹³ represented engagement as users clicked on the ad. For Twitter, reported “interactions”¹⁴ represented a similar concept. For the blog, we measured the number of e-mails received about the study referencing the blog post.

Effectiveness

To measure enrollment, we calculated the number of participants who completed the consent process and enrolled in the study for each recruitment approach. Participants self-identified how they were recruited. Time staff spent on study-related activities was tracked. Costs were calculated as follows. For costs involved in traditional recruitment, we calculated time spent on tasks and multiplied by a median Clinical Research Associate 1 salary and benefits ($25.23/hour). Costs spent on social media ads were those reported by social media companies when billed for the ads.

Procedure

In-person participant recruitment

Data for in-person recruitment methods were manually tracked by staff during recruitment events using a data collection spreadsheet.

Social media participant recruitment

Data for social media recruitment were assessed using the online reporting tools provided by Facebook and Twitter, as well as Google Analytics.

Time spent

Staff tracked the time spent during each activity and rounded to the nearest half hour.

Results

We recruited a total of 46 teens 14–18 years old in the Seattle area for a pilot physical activity intervention study. Participants had a mean age of 16 years (SD = 1.1); 57% were female and 28% nonwhite. The reported mean annual household income was $40–50,000, and median parent education level was some college/technical school.

In-person recruitment reached a total of 297 potential participants of which 220 engagements were logged, 90 were screened and eligible, and 37 enrolled in the study. Thus, the proportion of participants recruited represented 4.2% of participants who had some engagement with the research staff and 1.9% of the total potential audience. In-person recruitment required 28 hours of staff time. Thus, in-person recruitment required an average of 0.75 hours of staff time and cost an average of $19.09 per participant enrolled.

Social media recruitment reached a total of 34,272 potential participants of which 330 engagements were logged, 21 were screened and eligible, and 8 enrolled in the study. The proportion of participants recruited represented 2.7% of participants who had engagement with the advertisement and 0.026% of the total potential audience. Social media recruitment required 13 hours of staff time. Social media recruitment methods required an average of 1.6 hours of staff time and cost an average of $40.99 per participant enrolled. Table 1 compares these approaches.

Table 1.

Comparison of Recruitment Methods Between Social Media and Traditional In-Person

Recruitment method	Reach/approached	Ad click or flyer/engagement	Screened	Consented/enrolled	Hours of staff time	Cost per participant	Staff hours per participant
In-person school	275	200	86	35	17.5	$10.00	0.5
In-person clinic	22	20	4	2	10.5	$105.00	5.3
Total in-person	297	220	90	37	28	$15.14	0.75
Facebook ad version 1	20,904	91	4	1	3	$60.00	3
Facebook ad version 2	11,827	98	13	5	3	$12.00	0.6
Both Facebook ads	32,731	189	17	6	6	$20.00	1
Twitter	1,320	129	1	1	3	$60.00	3
Blog	221	12	3	1	4	$80.00	4
Total social media	34,272	330	21	8	13	$32.50	1.6

Discussion

This study utilized two approaches to recruit adolescents to a pilot intervention. We found that in-person approaches yielded participants in a narrow geographic area at a lower cost and staff time compared to social media. Social media approaches reached more potential participants from a wider geographic area and received more engagements compared to in-person approaches. Overall, we achieved diversity in participants recruited by gender, race/ethnicity, and socioeconomic status. While the cost per participant was higher for social media recruitment, these ads yielded more potential participants per staff hour invested. Furthermore, social media recruitment provided opportunities to reach youth across a wider catchment area, avoiding concerns with clustering of participants as occurs with school-based recruitment.

These findings add to the growing literature describing social media recruitment methods. Despite the nearly ubiquitous use of social media by adolescents, only one other published study describes social media research recruitment methods targeting adolescents. That study used targeted ads to identify teens with a rare genetic syndrome and found that technology was a more successful approach compared to in-person recruitment.¹⁵ Given that this previous study was focused on a rare genetic condition, social media may yield a more robust opportunity to identify and recruit targeted groups of participants. Our study used social media to recruit from a general population of adolescents and findings do not support in-person or social media as being clearly superior.

Placed in the context of the larger literature of studies using social media for recruitment, we found that social media recruitment costs about $430 per participant, while previous work with adult participants reported recruitment costs of $20 per participant. Contrary to expectations, this suggests that adolescents may be a more challenging population to recruit through social media than adults. While social media recruitment reached more teens than in-person recruitment, fewer signed up to be in the study. One possible explanation for this finding is that teens may have received many messages in schools or from parents about being wary to click on advertisements, while adults may be more trusting of clicking on an ad for a research study. While social media was more expensive per participant enrolled, it freed up staff time to engage in other study-related activities while online recruitment was in process.

Our study was limited in the small sample size we recruited, which represented our goal recruitment numbers for this pilot study. However, with this small sample size, we were unable to make meaningful comparisons between participants recruited through in-person compared to social media methods. Our recruitment approaches involved two different Facebook advertisements due to concerns raised by the IRB, which complicates our study process. However, as other researchers may also face concerns by the IRB about posting incentive promotions on social media, we hope that our findings will support their efforts. Third, the costs reported in this study are based on median staff salary in Seattle, WA. Other studies in different locations will undoubtedly have different calculated staff costs based on local salary norms.

Despite these limitations, our study has important implications. Researchers are increasingly called upon to apply innovation in their study designs, which may include encouragement to use social media to recruit. Our study findings suggest that the benefits and downsides of such approaches must be weighed for each study considering the population of interest and whether targeted approaches are needed. Future work should consider evaluating whether the characteristics of participants recruited in-person differ from those recruited through social media.

Footnotes

Acknowledgment

This study was funded by support from the Seattle Children's Research Institute.

Author Disclosure Statement

No competing financial interests exist.

References

Hooven

, Walsh

, Willgerodt

, et al. Increasing participation in prevention research: strategies for youths, parents and schools. Journal of Child and Adolescent Nursing, 2011; 24:137–149.

Dumka

, Garza

, Roosa

, et al. Recruitment and retention of high-risk families into a preventive parent training intervention. Journal of Primary Prevention, 1997; 18:25–39.

Moyer

, Finney

, Swearingen

, et al. Brief interventions for alcohol problems: a meta-analytic review of controlled investigations in treatment-seeking and non-treatment-seeking populations. Addiction, 2002; 97:279–292.

Lenhart

. (2015) Teens, social media & technology overview 2015. Washington, DC: Pew Internet and American Life Project.

Ramo

, Rodriguez

, Chavez

, et al. Facebook recruitment of young adult smokers for a cessation trial: methods, metrics, and lessons learned. Internet Interventions: the Application of Information Technology in Mental and Behavioural Health, 2014; 1:58–64.

Mitchell

, Petroll

. Patterns of HIV and sexually transmitted infection testing among men who have sex with men couples in the United States. Sexually Transmitted Diseases, 2012; 39:871–876.

Hernandez-Romieu

, Sullivan

, Sanchez

, et al. The comparability of men who have sex with men recruited from venue-time-space sampling and facebook: a cohort study. Journal of Medical Internet Research, Research Protocols, 2014; 3:e37.

Nelson

, Hughes

, Oakes

, et al. Estimation of geographic variation in human papillomavirus vaccine uptake in men and women: an online survey using facebook recruitment. Journal of Medical Internet Research, 2014; 16:e198.

Lohse

, Wamboldt

. Purposive facebook recruitment endows cost-effective nutrition education program evaluation. Journal of Medical Internet Research, Research Protocols, 2013; 2:e27.

10.

Pumper

, Mendoza

, Arseniev-Koehler

, et al. Using a Facebook group as an adjunct to a pilot mHealth physical activity intervention: a mixed methods approach. Studies in Health Technology and Informatics, 2015; 219:97–101.

11.

World Health Organization. Adolescent development. WHO identifies adolescence as the period in human growth and development that occurs after childhood and before adulthood, from ages 10 to 19. www.who.int/maternal_child_adolescent/topics/adolescence/dev/en (accessed Sep. 19, 2016).

12.

Google. (2016) Google analytics. www.google.com/analytics/#?modal_active=none (last accessed September 9, 2016).

13.

Facebook. (2016) Advertising on Facebook. www.facebook.com/business/products/ads/?campaign_id=1604904836426745 (last accessed September 9, 2016).

14.

Twitter. (2016) Advertising on Twitter. https://business.twitter.com/en/advertising.html (last accessed September 9, 2016).

15.

, Smaldone

, Fennoy

, et al. Using information technology and social networking for recruitment of research participants: experience from an exploratory study of pediatric Klinefelter syndrome. Journal of Medical Internet Research, 2013; 15:e48.