“The elephant in the room”: social responsibility in the production of sociogenomics research

Abstract

Sociogenomics examines the extent to which genetic differences between individuals relate to differences in social and economic behaviors and outcomes. The field evokes mixed reactions. For some, sociogenomics runs the risk of normalizing eugenic attitudes and legitimizing social inequalities. For others, sociogenomics brings the promise of more robust and nuanced understandings of human behavior. Regardless, a history of misuse and misapplication of genetics raises important questions about researchers’ social responsibilities. This paper draws on semi-structured interviews with sociogenomics researchers who investigate intelligence and educational attainment. It does so to understand how researcher’s motivations for engaging in a historically burdened field connect to their views on social responsibility and the challenges that come with it. In interviews, researchers highlighted the trade-off between engaging in socially contested research and the potential benefits their work poses to the social sciences and clinical research. They also highlighted the dilemmas of engaging with the public, including the existence of multiple publics. Finally, researchers elucidated uncertainties over what social responsibility is in practice and whether protecting against the misuse and misinterpretation of their research is wholly possible. This paper concludes by offering ways to address some of the challenges of social responsibility in the production of knowledge.

Introduction

Understanding who we are, our differences and similarities, and how they relate to social change, organization, and action is a critical aim of the social sciences. Social genomics, or sociogenomics, has recently started to offer new perspectives on this objective. Sociogenomics is an interdisciplinary field that explores “whether, how, and why genetic differences between human beings are linked to differences in behaviors and socioeconomic outcomes” (Harden and Koellinger 2020, p. 1). Using ever larger samples of genetic data, researchers examine social behaviors and outcomes like income (Hill et al. 2019), intelligence (Savage et al. 2018), and educational attainment (Lee et al. 2018).

The field invokes skepticism for some (Comfort 2018; Gillborn 2016; Specter 2014) and optimism for others (Harden and Koellinger 2020; Mills and Tropf 2020). This polarization (Adam 2019) is partially due to the historically burdened practice of viewing genetics in fatalistic and essentialist terms (Heine et al. 2017). Fundamentally, however, the fear of sociogenomics is rooted in the use of genetic ideologies to legitimize social inequalities (e.g., Herrnstein and Murray 1996; Jensen 1991), particularly those that fall along racial or socioeconomic lines. This history begs the question of whether research on a subject like intelligence could ever be socially neutral (Roberts 2015). Given the contested nature of sociogenomics in today’s (and previous) social contexts, this paper explores questions of social responsibility in the production of sociogenomics research. I begin by exploring sociogenomics researchers’ motivations, including their views on the potentially beneficial applications of their work. I then turn to researchers’ perspectives on their individual social responsibilities and the responsibilities of their field, focusing in particular on the dilemmas and challenges of public engagement. In doing so, I explore the relationship between researchers’ motivations for doing the work they do and their views on social responsibility. I apply a “tri-focal perspective” (Martschenko 2020) to researchers’ accounts, looking to the past, present, and future to contextualize their views. This paper concludes by providing recommendations for enabling greater emphasis of social responsibility in the production of knowledge.

The birth of sociogenomics

Although sociogenomics, as it is understood today, is a relatively new field, the term sociogenomics is not. Sociogenomics first appeared in the literature in the late twentieth century when molecular biologist Gene Robinson introduced the idea of identifying and understanding the genetic mechanisms influencing social behavior (Robinson 1999). While some of the earliest research in sociogenomics focused on insect populations (Robinson et al. 2005), current research in the field primarily investigates individual-level variations between humans within a single ancestral population.

Sociogenomics seeks to identify the genetic variants that are associated with human behaviors typically thought to be shaped by social interactions; it can be hard to separate from the field of behavioral genetics as there are overlapping analytical techniques, researchers, and stakeholders. Behavioral genetics traditionally includes clinical research on topics of interest to psychologists, such as psychiatric illnesses (Neumann et al. 2020), obsessive–compulsive disorder, cognitive ability (Malanchini et al. 2020), and ADHD (Ritter et al. 2017). Sociogenomics examines behaviors and outcomes that interest a larger cohort of social scientists—a community comprised of economists, political scientists, criminologists, and sociologists among others. Key topics of study in sociogenomics include income (Hill et al. 2019) and educational attainment (Lee et al. 2018); researchers are interested in their clinical relevance. For instance, scholars have explored the associations between polygenic risk scores for educational attainment and ADHD symptoms in children (de Zeeuw et al. 2014), and discussed the relationship between reproductive behavior and fertility traits such as childlessness or educational attainment (Mills et al. 2018).

Sociogenomics is governed by the ‘three laws of behaviour genetics’: (1) “all human behaviour is heritable”; (2) “the effect of being raised in the same family is smaller than the effect of genes”; and (3) “a substantial portion of the variation in complex human behavioural traits is not accounted for by the effects of genes or families” (Turkheimer 2000, p. 160). More recently a fourth law was instituted: “A typical human behavioural trait is associated with very many genetic variants, each of which accounts for a very small percentage of the behavioural variability” (Chabris et al. 2015, p. 305). This fourth law emerged as sociogenomicists began using genome-wide association studies (GWAS) to study complex traits.

GWAS are considered a valuable tool for investigating the genetic architecture of human illness and behavior (Bush and Moore 2012) and are used in both clinical and non-clinical research. GWAS incorporate individual-level genetic information from hundreds of thousands of individuals to identify single-nucleotide polymorphisms (SNPs) associated with a particular behavior or outcome; they produce a higher level of statistical power than the previous method of twin-studies (Polderman et al. 2015). The aggregate or combined effects of identified SNPs from across the genome form a polygenic score (PGS). A polygenic score can be constructed for any individual with genome-wide data and be used in later studies carried out in different samples. Importantly, GWAS identify correlations—the biological and environmental mechanisms through which a genetic variant becomes associated with a behavior or outcome are not commonly elucidated via GWAS. In other words, GWAS cannot reliably determine whether identified SNPs cause a particular behavior or outcome.

Sociogenomics also explores how environmental pathways moderate genetic influences. In doing so, it raises possible social, economic, and policy implications. For instance, in 2018 sociogenomics researchers published a study exploring the influence of genetics on social outcomes in Estonia during and after the Soviet era (Rimfeld et al. 2018). The authors concluded that genetic differences explained twice as much individual-level variation in social outcomes such as educational attainment and occupational status in the post-Soviet era than in the Soviet era (Rimfeld et al. 2018). The authors go so far as to suggest that social and policy changes, in this case the meritocratic selection of individuals in education and occupation that accompanied the transition from communism to capitalism, altered the extent of genetic influence in the sample population (Rimfeld et al. 2018). Other researchers have found that parents offer more cognitive stimulation to children with higher polygenic scores for education but that this pattern differs by socioeconomic status (Breinholt and Conley 2020); college-educated parents are less responsive to their children’s “genetic propensity towards educational success” than non-college-educated parents are (Breinholt and Conley 2020). In short, the period following the completion of the Human Genome Project in 2003 ended the nature-nurture debate. Today the conversation in fields like sociogenomics is a combination of nature and nurture (Martschenko 2020). Social and environmental contexts matter and can influence the extent to which a behavior or outcome is associated with genetic variants.

The proliferation of GWAS, the formation of large international research consortiums (e.g., Social Science Genetic Association Consortium, n.d.) and biobanks (e.g., UK Biobank, n.d.), and increasing media attention (Griffiths 2019; Harden 2018; Parens 2020; Specter 2014) showcase the growth of sociogenomics. Given the progress of the field and growing conversations about possible applications of sociogenomics research (Benjamin et al. 2012; Conley and Fletcher 2017; Harden and Koellinger 2020; Mills and Tropf 2020), this paper explores how sociogenomics researchers interested in intelligence and educational attainment identify their research motivations and think through their social responsibilities as scholars in a field rooted in the ugly history that the next section unpacks. In doing so, I hope to provide valuable context for starting a productive and multi-discursive dialogue that enables socially responsible communication of sociogenomics research objectives and findings.

Canary in a coal mine

Despite the promises of sociogenomics, research into the genetic etiologies of socially valued behaviors like intelligence are “fraught with social bias” (Roberts 2015, p. S50). Genetic ideologies about who is less or more ‘able’ or successful have been used to justify social hierarchies and explain away social inequality. For instance, the idea that there are genetic differences in educational ability has been used to paint educational equity a fruitless endeavor (Jensen 1991). In the United States, genetic ideologies were used to resist the abolition of slavery (Evrie 1868), advocate for more restrictive immigration policies (Brigham 1922), and involuntarily sterilize the ‘feeble-minded’ (Buck v. Bell 1927).

A central component of these arguments and practices are unfounded conceptualizations of race as a biological category instead of a social construction. African Americans were said to suffer from the ‘Negro IQ Deficit’ (Shockley 1971) making them less educationally capable than Whites (Beadie et al. 2017). This ‘molecularization of race’ (Fullwiley 2008) has historically been used to distract from social institutions and structures that perpetuate racial and socioeconomic inequality. Today, biological explanations for racial inequality persist (Wade 2014), raising concerns about emerging genetic technologies and the power they have to engender biologically deterministic reasoning about social hierarchies (Bliss 2018 2012; Duster 2005). In short, despite researchers best intentions, a field like sociogenomics has the potential to “function normatively to reinforce conceptions of race as an innate and immutable fact that produces racial inequalities” (Roberts and Rollins 2020, p. 196) because it offers a platform for others to sustain genetic explanations for racial differences in behaviors and outcomes (Panofsky 2018).

For these reasons, significant attention has been devoted to heightening social and ethical responsibility in social and behavioral genomics (Callier and Bonham 2015; The Hastings Center, n.d.). Research is a process of discovery that ideally supports human flourishing by ensuring that “technological innovation affords people freedom of expression and freedom from oppression” (Foley et al. 2016, p. 216). Sociogenomics encapsulates both ‘promise and peril’ (Bliss 2018), meaning it has the potential to support or stifle human flourishing. This dual-use dilemma (Selgelid 2009) posits genetics as “an object of fascination and fear” (Harden 2021, p. 1219). The risks, potential benefits, and ethical responsibilities of genetics research are core topics for scholars in the fields of bioethics (Parens et al. 2006) and ELSI (Ethical, Legal, and Social Implications), and entire academic research centers are devoted to these questions (National Human Genome Research Intitute, n.d.). Put simply, the genetic data revolution is underway. Alongside it is the canary in the coal mine—a growing conversation on whether and how genomics might stand in opposition to the principles of equity and justice, and expanding calls to proactively prevent warning signs of such outcomes from becoming reality.

Methods

This paper draws upon a larger mixed-methods study of teachers’ perceptions of intelligence, race and socioeconomic status in relation to genetics (Martschenko 2019). This study included interviews that took place between 2015 and 2016 with bioethicists and sociogenomics researchers. Here, I analyze interviews with sociogenomics researchers, focusing on their motivations and their perspectives on the role and relevance of sociogenomics research as well as their views on social responsibility given the, oftentimes, socially controversial nature of their work.

This paper incorporates semi-structured interviews with six sociogenomics researchers who focus on the relationship between genetics and intelligence or educational attainment. These researchers are highly educated, holding PhDs in fields like education, economics, and sociology. Four participants were from the United States and two from Europe. Five participants were male and one was female; a reflection of the underrepresentation of women in Science, Technology, Engineering, Mathematics, and Medicine workforce (Holman et al. 2018). All but one of the participants identified as White, a reflection of the highly White composition the field (Bliss 2018). Participants were purposefully selected because of their involvement in sociogenomics research on educational attainment and intelligence; two ‘traits’ that are highly socially valued and socially contested.

Six researchers is not a large sample. As such, this paper cannot make general claims about sociogenomics. However, this paper’s purposive sampling strategy and the relative size of the field of sociogenomics, and in particular the size of the community engaged in research on educational attainment and intelligence, provides a degree of moderatum generalizability (Payne and Williams 2005).

I approached potential interviewees as an applied bioethics researcher from a faculty of education. My background shaped my interest in the social and ethical implications of sociogenomics research and my focus on sociogenomics research on education-related behaviors like intelligence and educational-attainment. The semi-structured interview was selected over other forms of data collection because of the opportunity this method provides for reciprocity (Galletta 2013) and its versatility to responding to the interview environment (Kallio et al. 2016).

As the author of this paper and as the interviewer, I had no or limited prior rapport with interview participants, all of whom were initially contacted via email. In an effort to build rapport, participants were sent an interview guide ahead of time which was especially important given my positioning as an outsider to the field of sociogenomics. Participants consented to audio recordings of their interviews and were a given a transcript of their individual interview afterwards to review and to clarify their statements as they desired.

The semi-structured interviews started with questions about researcher motivations and background before transitioning to questions about sociogenomics research on education-related behaviors such as intelligence and educational attainment. Questions about the social, ethical, and policy implications of sociogenomics research, including the topic of social responsibility, were typically raised in the latter half of interviews, although their semi-structured nature means that these questions occurred during opportune moments rather than at fixed points. Interviews primarily took place over Skype video or phone and lasted approximately an hour. In some cases, informal discussion took place either prior to or after the interview. Finally, all participants have been renamed to preserve anonymity.

To make sense of the qualitative data generated via interviews, including ambiguities and interpretative challenges, this paper employs thematic analysis. Thematic analysis is a research tool for identifying patterns, or ‘themes’ across a data set (Clarke and Braun 2014). Braun and Clarke, (2006) present six-steps for carrying out qualitative thematic analysis: (1) familiarize yourself with the data; (2) generate initial codes; (3) search for themes among those codes to help with synthesis; (4) review themes; (5) define and name themes; (6) begin write-up. This paper approached qualitative analysis with an understanding that themes “reside in our heads from our thinking about our data and creating links as we understand them” (Anzul et al. 1997, pp. 205–206). Examples from the data that are included in this paper were selected in accordance with the study’s interest in researchers’ motivations and attitudes around social responsibility.

Themes were defined deductively, driven by an interest in understanding how a field such as sociogenomics came to be and how stakeholders in the field view their roles as producers of knowledge. Additionally, themes reside at the ‘latent’ level, meaning they go beyond the semantic content of interview transcriptions to “identify or examine the underlying ideas, assumptions, and conceptualizations–and ideologies–that are theorized as shaping or informing the semantic content of the data” (Braun and Clarke 2006, p. 13). This paper advances conceptual understandings of knowledge production in sociogenomics by moving beyond simple presentations of researchers’ motivations and views on social responsibility. In particular, it considers the ways in which researchers’ motivations influence their attitudes towards social responsibility. Additionally, I situate participant accounts within the wider academic and social context, constructing new ways to interrogate the production of knowledge in contentious fields like sociogenomics. At the same time, interview data do not purport to shed light on what participants truly feel; instead, the accounts shared serve as a window through which to explore the discursive construction of social responsibility in sociogenomics research.

Motivations for working in a historically burdened field

What drives the pursuit of knowledge in a field that evokes consistent discomfort, fear, and criticism? To varying degrees the researchers interviewed were motivated by the implications of their work for the social sciences, the practical applications of their research, and personal curiosity. They frequently contextualized their motivations by describing the current limitations and shortcomings of social science research and social institutions like education systems.

For instance, Carter and Grant described the possibilities genetic data pose for rigorous and robust research in the social sciences:

There’s an elephant in the room and that’s genetic inheritance. We know that obviously that’s part of what’s at play… We [social scientists] care about the things like class size [in education] or any number of units– things that we know are environmental effects. Because we’re not controlling for genetic differences, it’s going to be more difficult to find the true effects of those kinds of things. So, one idea is that measuring genotype provides a control variable that will allow better assessment, more precise assessment of environmental effects social scientists typically care about…It’ll eventually be standard to kind of control for some genetic genotype variable in order to make better [research] models better.

(Carter)

…I guess it’s that [GWAS results for educational attainment] was just kind of the elephant in the room, one couldn’t ignore it…In general society is going to be transformed by the availability of biological data across kind of all aspects of human life. I think the education system is going to have to cope with it, I think it’s going to be a salient feature of it. But my claim would not so much be that this is necessarily going to yield huge benefits… In terms of how it can change education research and social science research, I do think there’s some reasons to think that it could in certain ways improve the quality of research because it eliminates one clear confounder. There are lot of studies out that look at the association between variables and paint them in this plausible causal light and a clear confounder to a lot of those associations is that it’s the underlying biology driving the two traits—so to the extent that it could lead to sharper causal inferences and summaries, I think that would be useful in education research.

(Grant)

This idea—that genetic data could provide more nuanced understandings of social phenomena as explored through social sciences research—emerged repeatedly. It is worth situating this motivation within a larger discussion on the replicability crisis in fields like psychology (Romero 2019), where researchers interested in behavior genetics conducted candidate-gene studies that have since failed to replicate (Polderman et al. 2015). Replicability is grounds the epistemic authority of science– “we trust scientific findings because experiments repeated under the same conditions produce the same results” (Romero 2019, p. 1). As Grant and Carter cautiously spoke about the utility of genetic data, they presented it as an opportunity to gain a deeper, and perhaps more credible, understanding of human behavior and social outcomes. In other words, one motivation to emerge for conducting sociogenomics research was the belief that it offers an opportunity (whether legitimate or not) to ground the epistemic authority of the social sciences. This belief ties into the larger research enterprise, in which the replicability of GWAS findings has been described as “an unprecedented phenomenon” (Marigorta et al. 2018, p. 505).

The focus on what genetic data can provide to the social sciences as discussed by Carter and Grant is unsurprising. The accounts shared in this paper come from sociologists, education researchers, psychologists, and economists. Sociogenomics is a field that stands in a deeply interdisciplinary position—although its researchers are trained in genetics, it is not how many first began their careers (Bliss 2018). As such their goals may have shifted over time:

I was surprised by the influence of the biomedical perspective on nearly all health outcomes, and the extent to which the sociological perspective was pushed to the side. So, I wrote a K01 proposal to NIH in which I proposed to study statistical genetics and behavioral genetics in order to receive the necessary training to enter these debates…My goal was to remind researchers that people have friends, families, neighbors, co-workers—they have many social identities including gender, race, sexual orientation, political ideology, etc. And that it is nearly impossible to tell a single genetic story without considering these multiple social identities and roles…

(Eden)

Researchers also discussed their interest in the practical implications of sociogenomics research. For instance, Harold spoke about the potential clinical relevance of exploring genetic associations with cognitive ability:

When you identify variants that are associated with good or bad cognitive function, often the pathway, the biological or biochemical pathway that that gene is influencing reveals some aspect of how the brain works some aspect of neuroscience or some aspect of basic brain biochemistry. There are going to be enormous benefits from this kind of research for things like Alzheimer’s and schizophrenia and dementia. I think if you really sat down and thought about it, you’d realize for the amount of money that it costs to do these studies, compared to the ten or fifteen billion dollars that we spend on cutting edge high energy physics experiments like the large hadron collider at CERN, this stuff is ridiculously cheap for what you get back…it should just be kind of a no brainer that humans should work on this problem.

(Harold)

And, Olivia spoke about how genetics research might encourage education researchers and educators “to focus on individual differences” and nudge education to be more evidence based. Although she described genetics research on education outcomes as “still pretty primitive,” she went on to say:

…it’s [genetics research] developing really quickly, but it’s still quite primitive. I suspect we’ll look back in 25 years and think, ‘blimey we had these very blunt bits of information, and we were trying to apply them and what were we thinking?’ But I think it helps the general conversation that’s happening at the moment about making education more evidence based, so more like medicine…I think we can encourage that focus. And then I think what we can really do, is encourage a focus on individual differences, rather than on mean results, which is not one of the big forces in education we’re encouraging. The big international studies encourage this real focus on mean achievement, how we raise our national mean, or our schools mean, or our class’s mean. I think genetics research tells us to focus on individual differences and how we get the best out of every child and support every child in reaching their potential, learning as much as they can or finding what they’re good at and what they love.

(Olivia)

Harold and Olivia echo the assertions of Carter and Grant, highlighting the potential of genetics research to advance our knowledge of ourselves, our communities, and our societies. Harold and Olivia’s discussion of the practical applications of their work, whether in clinical settings or in classrooms, involves a re-imagination of current processes and institutions. Harold presents genetics research on cognitive ability as an opportunity to re-imagine how the research community thinks about Alzheimer’s, schizophrenia, or dementia. Olivia suggests that shifting focus from mean achievement to individual differences would be a departure from the conclusions of “big international studies” but ultimately enhance the credibility of education research and policies. These re-imaginations are grounded in the principle of research as a space for discovery. Research provides opportunities to learn and pursue curiosities. For Clive, these opportunities in and of themselves served as the initial motivation to conduct sociogenomics research on educational attainment:

I was driven just by curiosity. I mean the original goal was just pure curiosity. We wanted to know if we have such a gigantic sample [of individual-level genetic data], would we find something [genetic variants associated with educational attainment] or not. And that was a completely open question and if you would’ve asked me five years ago, I probably would’ve put my money on us just coming up with a gigantic no result even in a sample of 100,000. And this is basically how we started. We thought that it’s worthwhile to do this rigorously with a very large sample and even if we don’t find anything that—that is it still worth knowing. And then to our big surprise we actually found something–Oops.

(Clive)

Regardless of these researchers’ specific motivations, the themes of possibility and promise stood out, especially the notion that genetic data can enhance our study of social phenomena, social policies, and our health. These possibilities are raised at a moment in which there are discussions on whether there is a replicability crises in many academic fields (Fanelli 2018; Wingen et al. 2020) and environments continue to encounter longstanding social problems like educational inequality (Reardon 2016). This is not to say that genetic data will provide a solution to either. Indeed, a number of these scientists displayed a level of caution about the status of their field (e.g., “it is nearly impossible to tell a single genetic story”, or “my claim would not so much be that this is necessarily going to yield huge benefits”, or “it’s [sociogenomics] developing really quickly, but it’s still quite primitive”). Furthermore, there is legitimate fear that a field like sociogenomics might reify racism (Gillborn 2016) or open a new door to eugenics (Comfort 2018). Regardless, positioning researchers’ narratives within the wider academic and social context provides just as much insight into the appeal of sociogenomics for some as it does fear of it; research does not exist in a vacuum and neither does the motivation to (dis)engage in it.

The dilemmas of social responsibility

The display of caution exhibited by interview participants speaks to what Olivia describes as the “murky history” underpinning the field of genetics research. History informs science; this has significance for understanding the value of social responsibility and researchers’ perspectives on it in relation to their work. For example, the field’s “very murky history” and its “problems and fellows” make Olivia feel that she and her colleagues ought to “understand and accept that when we say certain things that might feel very benign, sometimes you get a very emotional response to them and that people are frightened.” She went on to say that the intense reactions the field can spark are justified: “…I think when people call us names, which does happen sometimes, there’s not an awful lot of point getting angry about it because it’s because of the history of our discipline.” Her lived experiences as a researcher interested in the genetics of education-related behaviors, and her recognition of the importance of the context in which knowledge is produced, encourage her to take her social responsibility seriously:

I feel very strongly that the onus is on us to educate people and to explain very carefully exactly what we mean when we talk about heritability and I think, and I know some people disagree with me on this, but I think we also have a responsibility to suggest, to try and interpret findings for the public and to suggest possible ways in which they could be used. I know there are people who think that shouldn’t happen, that a scientist should stop when the science is done, and hand it over to other people for the decisions, but I think in general, nobody understands what has happened quite as well as the people who’ve done it.

(Olivia)

Resnik and Elliott (2016) write that “recognizing one’s social responsibilities as a scientist is an important step towards exercising social responsibility, but it is only the beginning” (p. 1). Olivia certainly acknowledges her social responsibilities. Yet she also admits that social responsibility is not uniformly valued or exercised in her field.

Social responsibility is fundamental to the responsible conduct of research; structures like Institutional Review Boards (IRBs) are meant to protect the rights and welfare of research subjects and hold researchers accountable. In other words, responsibility on the part of the researcher is integral to the production of knowledge. However, what social responsibility looks like in the research context and how far a scientist should go to achieve it, is less clear cut (Beckwith and Huang 2005). Therefore while acknowledging one’s social responsibilities is an important step, “scientists may confront difficult value questions when deciding how to act responsibly” (Resnik and Elliott 2016, p. 1). There is no consensus on how to enact and embody social responsibility within scientific research. Does social responsibility stop at publication? Does it require direct engagement with the public? Olivia alluded to these dilemmas when she described “people who think… that a scientist should stop when the science is done.” The tensions Olivia draws out speak to several dilemmas in social responsibility: “1) dilemmas related to problem selection, 2) dilemmas related to publication and data sharing, and 3) dilemmas related to engaging society” (Resnik and Elliott 2016, p. 1). These dilemmas become especially complex in a historically burdened field like sociogenomics.

The researchers whose accounts are shared in this paper seemed most interested in the dilemmas in engaging society—whether and how to engage with the public. Clive described the time his team took “to think through the potential impact of the study…We were extremely concerned about how people may perceive this and also the potential of misinterpretation.” He also shared his experiences speaking with journalists in an effort to temper claims about his research on educational attainment:

We knew that this is extremely sensitive material from the very early days when we started discussing this type of work, and we had dozens if not hundreds of really serious conversations about this. People were really worried about social scientists starting to mingle with genetics and doing all sorts of crazy things. And I think a lot of the worries, they were totally warranted…We basically decided that we wanted to take a proactive approach to sort of make sure that things are not getting out of hand, that our words aren’t being twisted around and that people really understand what they can and cannot conclude from our results… the journalists want headlines that are breathtaking or sensational or, ‘gene for x has been found’. We spent a lot of time talking to journalists trying to convince them that they shouldn’t write such a paper or such a story and usually we considered it a success when after talking to us for an hour they decided not to write about it.

(Clive)

Eden echoed the important role of public media in communicating research results and seemed frustrated by misinterpretations of his work:

I have had 5–6 papers described in mainstream news outlets and I’m often surprised by their interpretations. A reporter might say “Researchers find a gene for smoking” which basically belies my entire point. There are no ‘genes for.” Genes and environments always co-occur and many believe that they co-evolve.

(Eden)

Although miscommunication and misinterpretation (e.g., “we had so many nightmares about this, we had so many worst-case scenarios playing through our minds while we were doing this study and preparing its publication.”—Clive) were central to how researchers thought about their responsibilities, Grant pointed out that there was no single audience to engage with. For him, engagement with peers in the social sciences was a more realistic exercise in social responsibility than tackling the general public:

…Shadows are definitely going to loom over the work that I do and so certainly shapes our work. I think we’re aware of that history and kind of doing our best to not fall into the same traps. By traps, I don’t mean problems of communication or interpretation. It is much more fundamental than that… There’s clearly a real problem in explaining scientific findings to the general public, attempting that is above my pay grade…I certainly feel some level of responsibility and the thing I feel a little more responsible for and maybe something I think I have a little more control over is trying to communicate with my peers…to the broader public that feels a little kind of far-field. There’s only so much I can do about that.

(Grant)

Grant points to the misleading idea that there is a single ‘general public’ to engage with. In reality, there are multiple publics and audiences; each might draw upon their lived experiences to interpret sociogenomics research differently. Grant saw value in focusing on his peers in the academy, perhaps because it might prevent other researchers, including “race researchers,” from misusing his work as a platform for biological determinism about racial inequality (Panofsky 2018). In comparison, Clive chose to engage with journalists. Regardless, both these audiences (journalists and researchers) are conveyors of information to the wider ‘publics.’ In this way, Grant and Clive navigate the dilemmas of engaging society through dialogue with communities who in turn may communicate with the wider public, or publics, directly.

Whether and how to engage with publics, is one question. Another is whether securing engagement with publics is enough to safeguard against the misinterpretation or misuse of research. Socially contested research might always cause anxiety or succumb to misinterpretation or misuse. Carter seemed to feel that while researchers should be conscious of their social responsibilities around the interpretation of their work, there was no fool-proof solution to the dilemma of engaging society:

I think you can do some things to control how your research is interpreted, but I don’t think you can ever fully control [how your research is interpreted] in the political sphere, and yes, it’s more dangerous for highly politicized topics, but I don’t think that’s a reason to avoid. I’m a big believer in the enlightenment—that you don’t restrict knowledge… you seek to understand things through scientific method whether or not they’re controversial or whether or not you may be comfortable with the answers or not. If we really just wanted to avoid a controversy, we would just not do this…You can do things yourself responsibly but you can’t control what other people do with it.

(Carter)

Furthermore, Harold stressed that despite the sensitivity of his research, its value outweighed possible risks:

It’s a sensitive topic, and it should be a sensitive topic because—and I don’t want to put all the blame on Europeans—but when in the recent history of colonialism and imperialism emanating mainly out of Europe, it was very easy for them to say ‘hey the reason we’re dominating these other people and exploiting these other people is because we’re better than they are, we’re smarter than they are and we deserve to dominate them’. So, there’s a whole intellectual terrain that’s highly contested here and so I understand that people are very sensitive about this topic and bad science has been used in the past in service of Empire and exploitation—I mean in the same way that guns have been used and rockets and atomic bombs have been used for various purposes. But it doesn’t mean that we shouldn’t develop those technologies.

(Harold)

Harold’s discussion on the utility of sociogenomics research relative to its controversy also touches upon social responsibility dilemmas related to publication and data sharing (Resnik and Elliott 2016) as well as the ethical dilemmas of dual-use research (Selgelid 2009). He references “guns” and “atomic bombs” as examples of other advancements that had adverse effects and yet were still worthy of development and incorporation into society. His perspective demonstrates the relationship between one’s motivations and their perspectives on social responsibility in the production of knowledge. Harold is motivated to conduct this research because of a perceived social good (e.g., “There are going to be enormous benefits from this kind of research for things like Alzheimer’s and schizophrenia and dementia”) that outweighs possible risk. His views on social responsibility and the dilemmas that surround it are shaped by how he justifies his work. Put another way, the onus of social responsibility, as Resnik and Elliott (2016) define it, does not seem to weigh heavy on Harold, even as he acknowledges that “there’s a whole intellectual terrain that’s highly contested here”; perhaps this is because he believes his greatest social responsibility is to advance knowledge through his work. Another possible interpretation is that it may be easier for Harold to view opposition to his work as emotional (e.g., “I understand that people are very sensitive to this topic”) because he has limited personal membership to a group that has historically and/or could in the future be harmed by sociogenomics.

Internal trade-offs between risk and benefit, and the nuances of the context in which a piece of research is produced and disseminated, contribute to the lack of consensus on social responsibility in the production of knowledge; they make social responsibility a difficult undertaking. Although highly important, social responsibility raises difficult questions about its practical translation. For these researchers the line in the sand where social responsibility starts and where it stops cannot be drawn easily. This difficulty may be informed by the fact that scientists can encounter public backlash and scrutiny or compromise their objectivity, or reputation for objectivity, during public engagement (Resnik and Elliott 2016). In other words, social responsibility entails a level of risk to the researcher that they may weigh against the risk of misapplication or misuse of their findings.

Discussion

While the complex history behind sociogenomics casts an uncertain future, Resnik and Elliott (2016) provide recommendations for dealing with the dilemmas of social responsibility. One recommendation is “collaborations with scholars who have some experience and expertise in ethics, politics, or public policy” which “may help scientists deal with the value implications of their work” (p. 11). There are examples of this already in sociogenomics, (e.g., The Hastings Center, n.d.; Social Science Genetic Association Consortium, n.d.) including attempts at ‘adversarial collaborations’ (Kahneman 2003; Martschenko et al. 2019), or joint research efforts that bring together individuals trained in different research traditions who carry strong disagreements (Martschenko et al. 2019). These collaborations are meant to “lever-age internal disagreements into a dialogue that can help inform the broader field” (Martschenko et al. 2019, p. 2).

Additionally, there is some evidence to suggest that community-driven research partnerships (Wing 2002) and responsible innovation frameworks (Balmer et al. 2016; Stilgoe et al. 2013) can promote the social responsibilities of researchers and their institutions by democratizing a field and breaking down silos (Charles et al. 2020). Although discourse on how best to operationalize social responsibility and the extent to which it should be pursued are likely to remain, more widely incorporating adversarial collaborations, community-driven research partnerships, and responsible innovation frameworks into the production of sociogenomics could offer practical opportunities to translate the commitment to social responsibility into action. Here, open and clear communication are central to success.

Grant and Carter both called genetic inheritance “the elephant in the room”. This turn of a phrase capures the discomfort genetics research brings out in many. Perhaps there is another dimension to this elephant– the dilemmas of social responsibility in a field whose foundational history has been used to perpetuate systemic racism and classism. Given the ugly history behind sociogenomics, the idea of social responsibility is particularly salient, as is understanding the motivations behind engaging in contested or ‘sensitive’ research. This paper found that while researchers acknowledged the ugly history of genetics research (e.g., “our discipline has a very murky history”—Olivia) and sought in some instances to actively mitigate the potential harms of their work (e.g., “we wanted to take a proactive approach …we spent a lot of time talking to journalists”—Clive), they highlighted the impossibility of completely protecting against misuse or misinterpretation (e.g., “…I don’t think you can ever fully control [how your research is interpreted] in the political sphere”—Carter) and explained that the potential for misinterpretation did not justify stopping research (e.g., “It doesn’t mean that we shouldn’t develop those technologies”—Harold). These discussions also highlight the existence of multiple audiences, or publics, who might engage with scientific research. Translating an acceptance of one’s social responsibility into effective action is not easy. However, interdisciplinary, multi-discursive research and community partnerships could open up windows to effectively operationalizing social responsibility within the production of knowledge.

Biography

Daphne Oluwaseun Martschenko is a postdoctoral Research Fellow at the Stanford Center for Biomedical Ethics. She holds a Ph.D. and MPhil in Education from the University of Cambridge. Her scholarship advocates for and facilitates cross-disciplinary research efforts that promote socially responsible communication of social and behavioral genomics research findings.