Facebook recruitment and the protection of human subjects

Abstract

Social and behavioral scientists increasingly use Facebook to recruit research participants. Given the everchanging social media landscape, it is important to consider the ethical principles of using such a strategy. The aims of this methodological article are to (a) examine Facebook recruitment in light of the ethical principles of the Belmont Report (respect for persons, beneficence, and justice), (b) describe ethical challenges that may be faced in Facebook recruitment, and (c) provide recommendations for researchers interested in adopting this recruitment method. Ethical challenges inherent in Facebook recruitment include: selecting subjects fairly, privacy, and data security. Overall, Facebook is a beneficial resource for recruiting participants into research; however, researchers need to be aware of their responsibility in protecting human subjects.

In order to access research participants, researchers are turning to innovative methods such as online recruitment by means of social media. Facebook, a social media website, has observed a shift in user demographic and usability over the years. A decade ago, Facebook use was limited to younger generations. However, as of early 2018, Facebook remains the most utilized social media platform in the United States; nearly 68% of adults report having a profile and 75% of these users report accessing their profile at least once per day (Pew Research Center, 2018b). Given Facebook’s proliferation throughout American society, researchers have increasingly used Facebook to recruit a variety of samples, such as, parent caregivers of children and teenagers with cancer (Akard, Wray, & Gilmer, 2015), long-term smokers (Carter-Harris, Bartlett Ellis, Warrick, & Rawl, 2016), and family caregivers (Herbell & Zauszniewski, 2018). A 2017 systematic review identified the benefits of Facebook recruitment compared to traditional recruitment methods as reduced costs, shorter recruitment periods, and improved participant selection of young and hard-to-reach populations. Disadvantages included that participants need internet access to participate and the overrepresentation of young, white women (Whitaker, Stevelink, & Fear, 2017).

Previous research has established the feasibility of using Facebook for recruitment, as well as for intervention delivery (Jones, Lacroix, & Nolte, 2015; Ling et al., 2018; Ramo, Thrul, Chavez, Delucchi, & Prochaska, 2015; Waring et al., 2018), although the challenge of recruiting diverse samples also was noted (Oesterle, Epstein, Haggerty, & Moreno, 2018). Researchers interested in learning more about recruiting using Facebook have access to a variety of primers (Carter-Harris, 2016; Valdez et al., 2014; Weiner, Puniello, Siracusa, & Crowley, 2017); however, it should be noted that advances in social media occur quickly and such articles are rapidly becoming outdated.

Although Facebook has been adapted as a recruitment tool, most research focuses on feasibility, costs, and ability to recruit diverse samples. Only recently have nurse researchers begun discussing methodological challenges faced in social media research (Arigo, Pagoto, Carter-Harris, Lillie, & Nebeker, 2018). Additional focus on the ethics and the protection of human subjects in an age of prominent Facebook and other social media platform use is needed. The need for such conversations is highlighted by the Cambridge Analytica controversy in which information on over 50 million Facebook users was used to create voter profiles (Ingber, 2018) and, more recently, a Facebook security breach (on September 28, 2018) in which user’s name, contact information, and profile details were obtained (Facebook, 2018). Therefore, this methodological article addresses a critical gap, challenging researchers to carefully consider research rigor and the protection of human subjects when using Facebook for research purposes. Moreover, the purpose of this article is to examine Facebook recruitment in light of the ethical principles of the Belmont Report (respect for persons, beneficence, and justice), describe ethical challenges that may be faced in Facebook recruitment, and provide recommendations for researchers interested in adopting Facebook recruitment methods.

Belmont Report

In 1979, the Belmont Report was issued by the National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, a group charged with establishing a code of research ethics. The Belmont Report has since informed the development of the Common Rule from the U.S. Department of Health and Human Services. The Common Rule (Department of Health and Human Services, 2016) outlines the Federal Policy for the Protection of Human Subjects. The Belmont Report outlines three ethical principles for researchers to adhere to: respect for persons, beneficence, and justice.

Respect for persons means that individuals are treated as autonomous agents and that individuals with diminished autonomy are protected (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979). An autonomous person, as defined by the Belmont report, is “an individual capable of deliberation about personal goals and of acting under the direction of such deliberation (p. 4).” Individuals with diminished autonomy may include prisoners or those with impaired cognition. The most common approach to assuring respect for persons is the use of informed consent. Informed consent refers to the process in which individuals understand that participating in research is voluntary and that they can stop their participation at any time. The informed consent process begins at the first contact with potential participants. Within the context of Facebook, respect for persons involves considering the process of recruitment, methods for selecting participants, and treating potential participants as autonomous agents.

Beneficence refers to treating individuals in an ethical manner by respecting their decisions, protecting them from harm, and securing their well-being (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979). The Belmont report notes that “the term ‘beneficence’ is often understood to cover acts of kindness or charity that go beyond strict obligation. In this document [the Belmont Report], beneficence is understood in a stronger sense, as an obligation” (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979, p. 4). Beneficence is typically expressed by two rules: 1) “do not harm” and 2) maximize potential benefits while minimizing potential harms. Often, the risk for human subjects is not the nature of the data but rather what the researcher does with the data. Within the context of Facebook, beneficence leads to a consideration of potential participants’ privacy and data security.

The ethical principle of justice asks, “who ought to receive the benefits of research and bear its burdens?” (National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research, 1979, p. 5). Justice provides moral requirements regarding the selection of subjects, specifically that there are fair procedures for participant selection. The Belmont Report challenges researchers to examine if subject selection is based on easy availability instead of reasons related to the research question. Within the context of Facebook, justice requires that researchers consider the reasons for not only selecting Facebook as the recruitment method, but also determine which modality within Facebook (i.e., advertisements or posts in groups) is best suited for recruitment.

Ethical challenges in Facebook research

Ethical challenges may not easily correspond to a single principle. Therefore, the ethical challenges have been outlined below.

Selection of Participants

Researchers can use multiple methods such as advertisements (i.e., boosted posts) or researchers can manually post in private and public groups to recruit participants through Facebook. In order to demonstrate the Belmont Report principle of respect for persons, careful consideration to the rationale behind method selection is needed. If using an advertisement, the advertisement is subject to specific format and content guidelines and will be reviewed by the Facebook Advertisement Team (Facebook, n.d. a). Facebook Advertisement Team does not approve advertisement wording that highlights participant characteristics. For example, a research study examining social support and self-management behaviors was recruiting young adults with inflammatory bowel disease (Kamp, Luo, Holmstrom, Given, & Wyatt, in press). The image advertisement stated “Are you between the ages of 18–29 and diagnosed with inflammatory bowel disease?” The advertisement was not approved because it contained prohibited content. Specifically, Facebook states that “ads must not contain content that asserts or implies personal attributes” (Facebook, n.d. b). This includes: race, ethnicity, religion, beliefs, age, sexual orientation, gender identity, disability, mental condition, and financial status. The Facebook advertisement team approved the recruitment message when it stated, “complete a survey to help researchers learn about how young adults manage inflammatory bowel disease.” Thus, be aware of content prohibited by Facebook advertisements and carefully consider how to reach a targeted population while adhering to these rules. Recommendations: review Facebook policies prior to developing recruitment messages especially if the Institutional Review Board (IRB) needs to review recruitment messages; start early discussions with the IRB regarding their role in internet recruitment messages.

Participants also can be recruited through Facebook groups, collections of people with similar interests who voluntarily become a group member. Groups can have various privacy settings including: public, closed, and secret. Each group has different rules regarding posting research recruitment messages (e.g., some groups the researcher may need to obtain approval from an admin). The ethical principle of justice requires researchers to consider the reasons for selecting support groups. For instance, a researcher may select three support groups with high membership. However, there may be differences between individuals who are members in larger groups verses smaller groups as well as differences between individuals who are willing to participate in groups compared to those who do not. Recommendations: Consider the rationale for selecting Facebook groups; check group policies prior to posting advertisements; if group policies are not clear, seek permission from group moderators before posting in any group.

Although there have been changes in the user demographics of social media, social media recruitment still presents as a challenge for less represented populations, such populations include older adults and individuals with a high school education or less (Pew Research Center, 2018a). It should be noted that Facebook is the most widely used social media platform, regardless of race and ethnicity (Pew Research Center, 2015). However, a noticeable downward trend exists when examining social media use across the age groups and educational levels in the United States (Pew Research Center, 2018a). On one end of the spectrum, 88% of 18–29 year old individuals use at least one form of social media daily (Pew Research Center, 2018b); and on the other end of the spectrum only about 34% of adults 65 and older report that they have ever used Facebook or Twitter (Anderson & Perrin, 2017). More older adults are using social media than ever before which can be attributed to the upticks in tablet use, smartphone ownership, and internet access and affordability (Anderson & Perrin, 2017). Consequently, some researchers have experienced successes in recruiting older adults by means of social media. For example, Cowie and Gurney (2018) used Facebook to recruit healthy elderly people into a clinical trial and Musil (2018) uses Facebook to recruit grandmothers raising grandchildren. Researchers interested in recruiting older adults can review and adapt the National Institute of Aging’s Recruiting Older Adults into Research (ROAR) toolkit recommendations for use in social media recruitment (National Institute of Aging, 2015). While we acknowledge that additional research is needed to determine best practices in recruiting diverse populations on Facebook, researchers can begin to address this gap by reporting similarities and differences between the Facebook sample and population estimates. For example, Weiner and colleagues (2017) included population estimates to discuss participant self-selection, representativeness, and sampling population undercoverage among a sample of individuals over 50 years old recruited from Facebook.

Participant and Researcher Privacy

Issues of privacy relate to both respect for persons and beneficence. Privacy refers to an individual’s right to control access to personal information (Association of Internet Researchers, 2012). Within a social media context, privacy often has implications for both participants and researchers. Although Facebook users can control privacy settings, differences may exist in beliefs regarding public verses private data (known as perceived privacy; Association of Internet Researchers, 2012). In some instances, participants can make comments and “tag” (i.e., identify individuals by name) others on Facebook advertisements. This may become problematic for individuals who are trying to keep their disease state private such as an individual diagnosed with HIV/AIDS. If the individual with HIV/AIDS is tagged, the general public, friends, and family may be able to see that their loved one was tagged on a post related to a disease state. The comment feature is dependent on the type of ad selected. Although ads are presented for a select period of time, the use of advertisements should involve a consideration of the type of ad, length of recruitment time period, and potential effects of commenting.

In addition, online participants (e.g., groups) may be unaware that they are being monitored and do not have control over the type of data collected. One form of data extraction, with questionable legality, is called data scraping. Data scraping is when a computer program extracts publicly available data without user’s knowledge for another purpose. Facebook has been particularly susceptible to data scraping which was the primary approach of the 2018 Cambridge Analytica data collection and subsequent scandal that affected as many as 87 million Facebook users (Ingber, 2018). Recommendations: consider how privacy settings may be influenced by Facebook policies and procedures and therefore recognize that the researcher may not have control of privacy settings. Be clear with participants regarding data collected for recruitment purposes.

Privacy implications can also affect the researcher. For instance, posting in a support group involves making a post from a personal profile. Depending on the researcher’s personal privacy settings, potential research participants may be able to see personal photos or status updates that the researcher has posted. Potential participants also may be able to see where you live or work which may present as a safety issue regardless of the population of interest. Recommendations: be vigilant with personal Facebook privacy settings; read and review the lengthy updates about Facebook privacy that may change frequently. We also highly recommend reviewing your state’s Nurse Practice Act and the National Council of State Board of Nursing (NCSBN) social media guidelines (2018).

Data security

Though no part of the Belmont Report’s directives specifically extend to data security in a research recruitment context, data security is a universal protection for human subjects. Data security refers to the researcher’s responsibility to prevent the disclosure or loss of sensitive information collected during a research study (National Institutes of Health, 2018). When using Facebook for participant recruitment, researchers often collect general data, such as number of participants liking a post or advertisement, which typically do not fall into the category of Protected Health Information. In some circumstances, researchers may also collect identifiable participant information. The data security principles presented below can easily be extended to researchers using Facebook for data collection and intervention delivery. Careful consideration of the data collected during Facebook recruitment is warranted.

Even in online contexts, researchers are still required to comply with the Health Insurance Portability and Accountability Act’s (HIPAA) Privacy and Security Rules. Though both rules deal with protecting patients’ and research subjects’ Protected Health Information (PHI), the Security Rule is more specific. It requires the researcher to provide the “appropriate administrative, technical, and physical safeguards to protect the privacy of protected health information” for all PHI that is stored or transmitted through electronic means (ePHI) (Department of Health and Human Services, 2017). In other words, if the researcher is required to ensure that research subjects’ PHI is protected by placing it in a locked cabinet and ensuring only those that need to see that record have access, the Security Rule requires the same level of security (e.g., data are secured and access is limited) for records stored electronically. Even if Facebook recruitment data are not considered ePHI under HIPPA, for maximum data security, the PHI Security Rule should be applied. At its simplest, the Security Rule requires safeguards that are contingent to the dangers of improper disclosure facing that specific record holder. The safeguards must therefore be “reasonable” in light of the recognized dangers to that particular record holder.

Researchers working in healthcare must be aware of several factors that will affect what security measures they will be required to implement for data protection. The first factor is that recent trends indicate healthcare industries have become larger targets for hacking attempts (McCoy & Perlis, 2018). Cloud computing—when the researcher stores any relevant data in a “cloud,” which means a copy of the file is stored on servers other than the one the researcher may have physical access to—has also become a larger target of hacking (Giles, 2018). Moreover, while companies like Facebook do provide some data security measures, they are not required to abide by the same data security practices that are required by HIPAA and IRB guidelines. Indeed, as mentioned above, the recent Cambridge Analytica data breach has led to decreased confidence in Facebook’s ability to guarantee that data will remain in proper hands. Reasonable security measures by the researcher must take all three of these factors into account.

By taking these factors into consideration, the researcher can implement data security measures that will work for both researcher and research subject alike while staying in compliance with the HIPAA Security Rule and IRB guidelines. Our recommendation for data security is dependent on where the researcher plans on storing the recruitment information gathered through Facebook. Regardless of data storage location on a cloud service or a server owned by the researcher, integrity of access to those records is absolutely required. The first recommendation we have for data security is implementation of a multi-factor authentication process for access to those records. This can generally be done through a third-party, multi-factor authentication service which generates an alpha-numeric code. Thus, an individual attempting to access research data must input two passwords, a user created password as well as the alpha-numeric code provided by the app. This allows the information stored on the computer to have multiple layers of encryption protecting it against unauthorized intrusion. Generally, the researcher’s institution will have access to a multi-factor authentication app that can be used. Second, the researcher may provide even greater safeguards by simply requiring those with access to have strong, randomized alpha-numeric passwords that change on a regular basis. Finally, the researcher may continually monitor server access to ensure that unauthorized parties are not accessing the records.

Additional Considerations

There are additional considerations for researchers using Facebook as a recruitment resource. Ensuring anonymity in social media research is near impossible; therefore, anonymity should not be promised. Instead, researchers can focus on enhancing confidentiality through privacy and data security methods. A case study highlights the difficulty of anonymity in Facebook research. Researchers publicly released data collected on Facebook from the “Tastes, Ties, and Time” project. Although the data were de-identified, the dataset contained enough specific data to enable re-identification by an outside researcher (Zimmer, 2010).

Additionally, researchers need to be transparent with how, when, and where researchers are monitoring, interacting, and assuring security with participants. This transparency should be articulated in the informed consent but also addressed in any communication between the researcher and Facebook entity. A simple disclaimer may be placed at the bottom of the post stating that any comments or interactions with the post may be monitored by the researcher.

The landscape of participant recruitment for research is everchanging, thus, innovative recruitment methods must be utilized by researchers. While many researchers are successfully recruiting participants through Facebook, the scientific community must to be cognizant that protecting human subjects in an era of Facebook recruitment requires ongoing assessment. The Belmont Principles of respect for persons, beneficence, and justice can be used as a framework to examine Facebook recruitment methods and the protection of human subjects. As Facebook policies and procedures change, researchers should stay abreast of developments that may influence the ethical dimensions of recruitment efforts using Facebook.