P3B277: Is It Spam?: The Challenges of Using Online Advertisements to Recruit Prospective Parents to Participate in an Online Educational Intervention

Background: Online recruitment methods using targeted advertisements and social media posts are increasingly being employed to recruit study participants because of their vast reach and cost-effectiveness compared to in-person recruitment. Such methods became increasingly useful during the COVID-19 pandemic when in-person recruitment was restricted by many institutions. However, without face-to-face contact, it is difficult to know if responses are spam, with individuals or computer bots able to repeatedly enroll and potentially affect study results. We aim to assess the number of duplicate spam responses from potential study participants recruited via online advertisements on several commonly used social media platforms.

Methods: Online advertisements were placed on CraigsList, Facebook and Reddit platforms, seeking pregnant persons 24-32 weeks to participate in an online study educating prospective and new parents about breastfeeding, infant care and safety. Interested participants completed an eligibility form through Qualtrics. Information collected included contact information, name, date of birth, due date, whether carrying multiples and whether the individual expected to have custody of the infant after birth. Survey results and survey collection data provided by Qualtrics were then searched for duplicate entries. Surveys with duplicated information were analyzed for consistency in responses and were coded as either likely legitimate, with consistent contact and eligibility answers, or likely spam, with inconsistent contact information and survey responses from the same source. Spam duplicates were then sorted by platform placement and advertisement targeting, to see which methods yielded the least spam.

Results: Between November 2021 and October 2022 a total of 524 eligibility surveys were completed in response to online advertisements. 50 responses (9.7%) from a repeated IP address and 141 (27%) from identical latitude and longitude (L&L) points and were deemed likely spam based on discordant survey responses. An additional 39 responses came from repeat IP addresses but were deemed likely legitimate with consistent contact information and survey responses. Of the legitimate duplicates 21 (53.8%) had identical L&L points. Neither Qualtrics nor the any of the online social media platforms recognized survey respondents as suspicious or spam. The number of spam duplicates decreased across all platforms with more specific targeting of ad placement.

Conclusion: Online advertising is an effective way to recruit pregnant persons for social media and survey-based research. However, it is likely to yield a large proportion of duplicate entries with discordant data consistent with dishonest eligibility responses, which are unlikely to be identified as spam automatically. IP addresses and L&L of survey completion can identify repeated and discordant responses and is likely an effective initial screening tool. However, further work is needed to help researchers quickly identify bot and scam responses. Moving forward, researchers using online recruitment should disclose the methods used to identify spam respondents recruited via online advertisements in their methods.