Cybervetting, or as we discuss more generally, using social media information to assess applicants, is on the rise: organizational use of social media assessment increased by 20% in a 2-year period (2010–2012; Winter, Reference Winter2013). According to the Society for Human Resource Management (SHRM; 2011), 26% of human resource management (HRM) professionals use social media to assess person–organization fit during selection; more recently, that number has climbed to 43% (Maurer, Reference Maurer2016), and according to SHRM (2017), 84% of organizations report using social media for talent acquisition. Yet, in a review by Van Iddekinge etal. (Reference Van Iddekinge, Lanivich, Roth and Junco2016), findings show that recruiter ratings of social media information are generally unrelated to job performance, turnover intentions, and turnover. Thus, there is tension between good measurement practices in industrial-organizational (I-O) psychology and HRM, and the largely uncharted waters of using social media information to assess applicants.
In response to Wilcox etal. (Reference Wilcox, Damarin and McDonald2022), we argue first that the term “cybervetting” may be too narrow to include the multitude of ways this practice is examined in the literature and thus describe additional and related research in HRM that examines the same practices using different terms. In service of including more general research on this topic, we broaden our discussion from cybervetting to using social media more generally to assess applicants. Second, the focal article begins the important conversation regarding identity-related implications for individuals who post on social media (e.g., impression management). However, we take this conversation further to explicitly examine the tension that could exist for individuals with minoritized social identities (e.g., an individual with a certain social identity belonging to a group that has historically been marginalized or devalued) who engage in identity management (e.g., disclosure) online. If this involuntary information provided by people from protected classes is used for assessment purposes in selection, this may have very real intrapersonal consequences for those individuals regarding portraying their authentic, true selves online and for organizational representatives, who may be relying on out-of-bounds information. Finally, we address the validity-related implications of an unstructured, unformatted process when using social media information to assess applicants (as do Wilcox etal., Reference Wilcox, Damarin and McDonald2022), but we extend this conversation to explicitly address the additional adverse impact considerations, that in this case, without established validity evidence, may be even more paramount. In summary, although we understand the ubiquity of using social media information to assess applicants, we emphasize the lack of empirical evidence for this unstructured practice and additional measurement-related threats that arise from using it and conclude with methods-based guidelines for future research that should guide its potential utility.
Cybervetting or using social media information for assessment
Social networking sites (SNS; Pike etal., Reference Pike, Bateman and Butler2018) or social networking websites (SNW; Hartwell & Campion, Reference Hartwell and Campion2020) are the “socially connected, web-platformed, open information systems” that allow users to share and source information (p. 730). Because of the inherent ability of these systems, they allow information to be searched and sourced by a variety of audiences for whom the information may not have been intended (Ellison etal., Reference Ellison, Steinfield and Lampe2011). Research highlights the many reasons these information sources are used by hiring professionals, one of which is to strive to reduce ambiguity when selecting job applicants. Importantly, work examining these practices is not limited to the term “cybervetting” but more broadly referred to in the literature as social media information (Van Iddekinge etal., Reference Van Iddekinge, Lanivich, Roth and Junco2016), social media assessments (SMA; Roth etal., Reference Roth, Bobko, Van Iddekinge and Thatcher2016), and using social media content to screen or assess applicants (Wade etal., Reference Wade, Roth, Thatcher and Dinger2020). Critically, we must include these related if not more general constructs from HRM and organizational behavior (OB) in an examination of this space. Due to construct proliferation, these terms have been used interchangeably in the literature and importantly must be included in a comprehensive discussion of this research base. As these additional sources discuss, using social media information to assess applicants during the hiring and selection process (the general term we will use to describe this process here) can cause incongruences and actually increase ambiguity (Pike etal., Reference Pike, Bateman and Butler2018). Because hiring personnel examine a variety of social media information sources (e.g., Facebook, LinkedIn, Twitter), they often flatten the information gleaned across multiple audiences into one source of information (e.g., personal, professional; friends, family; this is called “context collapsing”; Marwick & Boyd, Reference Marwick and Boyd2011, p. 122). The expectation to present a singular identity makes it impossible to differentiate self-presentation strategies for different sites (e.g., personal, professional), creating tension as individuals decide what is posted where and the privacy settings of the post as large groups of diverse people begin information sourcing from social media (Boyd, Reference Boyd2008). The tension among type, completeness, and accuracy of the information posted in various places online raises questions regarding how applicants choose to impression manage (Duffy & Chan, Reference Duffy and Chan2019) online and across sites, which may be particularly important and relevant if they have a protected social identity, privacy settings enacted, and an expectation of privacy if they choose to disclose online. This tension also highlights questions regarding the validity of using these assessments (which Wilcox etal. [Reference Wilcox, Damarin and McDonald2022] raise in the focal article) and, importantly, because the validity of using social media information to screen or assess applicants has rightfully been questioned, the subsequent adverse impact of doing so.
Online impression management and implications for identity management
People of minoritized social identities may use social media to share or seek social support specifically related to their belonging to one or more marginalized social identities (e.g., mothers, veterans, stigmatized national origin, LGBTQ+ communities). Specifically, research documents that individuals who identify as LGB use social media more frequently than those who identify as cisgender (Escobar-Viera etal., Reference Escobar-Viera, Shensa, Bowman, Sidani, Knight, James and Primack2018) and that social media has been a critical harbinger for social change in terms of gender equity activists across college campuses (Linder etal., Reference Linder, Myers, Riggle and Lacy2016). Yet, there may be tension between presenting oneself as one wishes to be perceived (e.g., impression management: fostering impressions of oneself in others’ eyes; Leary & Kowalski, Reference Leary and Kowalski1990) and seeking instrumental or psychosocial support related to one’s social identity online (e.g., identity management: how individuals manage invisible social identities in the workplace; Clair etal. Reference Clair, Beatty and Maclean2005). One example of this tension manifested is the popularized “selfie”; Barker and Rodriguez (Reference Barker and Rodriguez2019) found that participants take selfies to “say something about who they are, connect with others, feel better about themselves, feel empowered, and … identify with others like themselves” (p. 1158). The authors found that individuals who took and posted selfies were motivated by wanting to feel better and more empowered. Specifically, within marginalized social identities, selfies or, even more broadly, being oneself online can provide a safe environment that provides an individual with the freedom to connect with others who may belong to the same race or sexual orientation (Barker & Rodriguez, Reference Barker and Rodriguez2019). Numerous studies have shown that people benefit both mentally and physically from expressing and defining themselves in complex, multifaceted ways (Shaw & Gant, Reference Shaw and Gant2002). Providing a safe environment within a group of “like-minded” individuals can be an important aspect of pride that allows free expression and safety to voice one’s personal opinions in what may be considered a vulnerable and risky situation. For example, individuals of minoritized social identities may find it challenging to present who they are for fear of being discriminated against, outcast, judged, or mistreated. But a more authentic online presence may be used to gain affirmation from others who identify similarly (Barker & Rodriguez, Reference Barker and Rodriguez2019).
This tension between online impression management (e.g., presenting one’s best self) and online identity management for people of minoritized, invisible (or at least unobservable) social identities (e.g., presenting one’s true self; Berkelaar, Reference Berkelaar2014) prevents job applicants from being able to control their own narrative due to the discoverability of this information when social media is used to assess applicants. In addition to these threats to the self, for protected, minoritized social identities in the United States (e.g., pregnancy, disability, ethnicity), it is illegal for employers to ask about these identities during the hiring process, much less make hiring decisions based upon them. This may create a host of issues for employers when it comes to unregulated use of social media information to assess applicants.
Importance of valid and structured practices in selection
One reason hiring personnel or agents use social media information to assess applicants is an intention to increase efficiency through the assessment of candidate fit. Hiring personnel use social media information as a means of risk reduction (through due diligence and professional identity work) and reputation management (Berkelaar & Buzzanell, Reference Berkelaar and Buzzanell2014). However, it is empirically unsubstantiated that using social media information in selection is significantly related to important hiring outcomes such as job performance. For instance, researchers have found that social media content, personality, and job performance have relatively low correlations (e.g., Kluemper etal., Reference Kluemper, Rosen and Mossholder2012; Van Iddekinge etal., Reference Van Iddekinge, Lanivich, Roth and Junco2016), subsequently raising concerns over faulty inferences and overinterpretation of data retrieved from social media. For example, applicants who otherwise fit hiring criteria have been rejected due to hiring managers seeing their social media posts related to alcohol consumption (Davison etal., Reference Davison, Maraist and Bing2011). Additionally, Wade etal. (Reference Wade, Roth, Thatcher and Dinger2020) found that hiring managers have a bias toward hiring those who are politically similar to themselves, specifically regarding marijuana legalization, gun control legislation, and health care affordability. Using social media information to assess applicants remains under researched, as Wilcox etal. (Reference Wilcox, Damarin and McDonald2022) note, and it is not currently understood whether this wealth of information is actually beneficial (Roth etal., Reference Roth, Bobko, Van Iddekinge and Thatcher2016).
To justify using social media information when assessing applicants, it is important to establish valid and structured practices that predict job performance. Aselection process is considered valid if it increases the employers’ chances of hiring the right person for the job (Billikopf, Reference Billikopf2003). This may include establishing and using indicators for valued outcomes (such as job performance) during evaluations or hiring decisions (Bauer etal., Reference Bauer, Erdogan, Caughlin and Truxillo2020). Additionally, valid measures should be reliable such that they perform consistently when reproduced under the same conditions.
The lack of structure when using social media information to assess applicants can be highly problematic. Hiring personnel are not using clear and consistent evaluation criteria, which may lead to inconsistent implementation across raters and between job candidates. Social media (and “Googling”) allows greater access to information such as race, age, and other characteristics that could increase potential discrimination (Sharone, Reference Sharone and Vallas2017) of protected groups. This bias and unfairness may consequently disadvantage underrepresented groups (Ruggs etal., Reference Ruggs, Walker, Blanchard, Gur, Landers and Schmidt2016). Also, hiring personnel may not focus on job-relevant criteria, which makes it more challenging to connect to desired outcomes such as job performance or the knowledge, skills, abilities, or other characteristics (KSAOs) listed in the job description (Bauer etal., Reference Bauer, Erdogan, Caughlin and Truxillo2020). By using information that is unrelated to the job for selection, hiring agents are susceptible not only to selecting the wrong candidates for the position but also to creating liability.
Implications for adverse impact and protected groups
Because employers are justifiably concerned that traditional selection methods (résumés, cover letters, and interviews) are vulnerable to deception via impression management tactics, which could exaggerate or obscure job-relevant KSAOs, using social media information to assess applicants provides a potential solution to balance hiring the “right” person while avoiding negligent hiring lawsuits (Berkelaar, Reference Berkelaar2014; Vosen, Reference Vosen2021). However, lacking structure or regulation, using social media information to assess applicants may blur ethical and legal lines regarding privacy while indirectly facilitating employment discrimination, as protected statuses may be observed online (Hoek etal., Reference Hoek, O’Kane and McCracken2016; SHRM, 2016; Vosen, Reference Vosen2021). Stamper (Reference Stamper2010) found that 45% of 2,600 U.S. hiring managers used social media information in selection and that 35% of this group rejected candidates based on the information yielded from social media. The ethical and legal justifications for these rejections remain inconclusive. Furthermore, reports from a senior manager at the Equal Employment Opportunity Commission noted that “approximately 75% of recruiters are required to do online research of applicants, and 70% of recruiters surveyed reported rejecting individuals as a result” (Roth etal., Reference Roth, Bobko, Van Iddekinge and Thatcher2016, p. 270). However, Title VII prohibits basing employment decisions on protected status unless it is a bona fide occupational qualification (e.g., a church requiring ministers to be Christian or members of the denomination). Nevertheless, some employers continue to make decisions on such factors, and this can even be more pronounced when using social media information to assess applicants (Hebl etal., Reference Hebl, Cheng and Ng2020; Jones etal., Reference Jones, Arena, Nittrouer, Alonso and Lindsey2017). Specifically, attraction-selection-attrition theory explains this relationship by characterizing organizations that tend to attract, select, and retain individuals who possess similar attributes (e.g., personal characteristics, values, and interests) to existing organizational members (Schneider, Reference Schneider1987).
Using social media information to assess applicants also highlights the digital divide, how the use of and access to the internet varies across demographics such as age, gender, and race, which could perpetuate hiring discrimination (Alexander etal., Reference Alexander, Mader and Mader2019; Roth etal., Reference Roth, Bobko, Van Iddekinge and Thatcher2016; Vosen, Reference Vosen2021). Popular press suggests that applicants with a social media presence may be viewed more positively than those without because the absence of online information may lead to uncertainty about certain attributes or qualifications of an applicant (Hill, Reference Hill2012). Research shows that social media is used to evaluate competencies such as creativity, cognitive ability, and interpersonal skills, which are essential skill sets for most jobs (Alexander etal., Reference Alexander, Mader and Mader2019; Roth etal., Reference Roth, Bobko, Van Iddekinge and Thatcher2016). Furthermore, a survey examining social media use in hiring and job seeking found that younger applicants and human resource professionals tended toward Facebook and job boards in job seeking, that a higher education level is associated more with LinkedIn for job seeking, and that men tended to use LinkedIn more for both hiring and job seeking than do women (Alexander etal., Reference Alexander, Mader and Mader2019). According to an article in Forbes, women tend to rely more on personal networks of friends and family in job seeking (Huang, Reference Huang2017). Given these variations in social media use, it is plausible that older people, women, and less-educated individuals could be adversely affected when social media information is used to assess applicants.
Jacobson and Gruzd (Reference Jacobson and Gruzd2020) posit that social media screening has introduced new forms of discrimination, such as evaluating a candidate’s influence based on the size of their social network and using photo scanners to predict the mental well-being of a candidate. Algorithms have been designed to predict with high accuracy whether someone is suffering from a mental disorder such as depression, suicide ideation, and schizophrenia (Chancellor etal., Reference Chancellor, Baumer and De Choudhury2019; Owusu etal., Reference Owusu, Reininghaus, Koppe, Dankwa-Mullan and Bärnighausen2021). Such advanced algorithms can be used to evaluate mental health based on image characteristics such as colorfulness, sharpness, naturalness, and facial presentation (Owusu etal., Reference Owusu, Reininghaus, Koppe, Dankwa-Mullan and Bärnighausen2021). Although these efforts are laudable because they drive public health efforts to understand and design interventions for mental health disorders, the use of such algorithms in using social media information to assess applicants could promote selection biases and Type Ierrors (concluding falsely that a candidate is mentally ill when they are not). It is well documented in literature from I-O psychology and organizational behavior that individuals with mental health disorders hesitate to disclose such identities for fear of being stigmatized and limiting their employment opportunities (Toth & Dewa, Reference Toth and Dewa2014). Indeed, a study of whether employers were less likely to hire an individual with a mental health problem revealed that previously depressed candidates were less likely to be recommended for hiring than those with no disability and that there was a significant difference in employers’ attitudes toward employing people with mental disabilities and physical disabilities such that the former was viewed more negatively because of the general stigma attached to mental illness (Brohan etal., Reference Brohan, Hendserson, Wheat, Malcolm, Clement, Barley, Slade and Thornicroft2012).
Given these concerns for equity, using social media information to assess applicants may not only be invalid (particularly without a structured approach to assessment) but also induce adverse impact. Generating valid and structured practices for using social media information to assess applicants is the best way to be able to consistently measure their implementation and reduce their adverse impact.
Intentional, data-based practices are critical
Importantly, as Wilcox etal. (Reference Wilcox, Damarin and McDonald2022) describe, if hiring personnel or agents are largely ambivalent about using social media information to assess applicants, they should be persuadable regarding the practices they use if they must do so. We argue here that looking outside of the cybervetting literature alone is an important first step, as previous work has been conducted to empirically examine the (lack of) evidence regarding the validity of using social media information to assess applicants. Second, we make the case that in broadening this definition, we also need to dive deeper regarding the potential negative identity-management implications for both the self as well as organizational-level outcomes when hiring personnel use involuntary information that may be shared online by minoritized applicants when conducting these hiring assessments or making inferences due to a lack of online presence. Third, Wilcox etal. described the critical role of a valid process if using information provided on social media for assessment purposes, and we extend that argument to an additional repercussion of not having this validity evidence: mainly, increasing the threat of adverse impact during selection.
In conclusion, given the prevalence of social media usage and its use for assessing applicants by hiring personnel, we advise that recruiters, hiring personnel, or hiring agents who want to use this information in their assessments wait until established validity evidence supports this practice or, minimally, to proceed with caution. Without using a wide-angle lens that incorporates the full scope of this literature or taking into account the ways inconsistent implementation may systematically disadvantage minoritized job applicants, there are substantial risks inherent in a laissez-faire approach to this practice. If social media information must be used, which we currently argue against, at a minimum, establishing a structured practice, with clear operational definitions and a consistent approach to information gathering, would be a crucial first step to possibly generating some evidence for a replicable process. Wilcox etal. (Reference Wilcox, Damarin and McDonald2022) establish some first-look guidelines, but much more work is required to empirically substantiate an approach that is valid but also does not increase adverse impact. Without discussion and empirical vetting of these important trade-offs, unfettered use of this practice is ill-advised.
Open access
