Performing Trustworthiness: The ‘Credibility Work’ of Prominent Sociologists

Abstract

To the limited extent that sociologists have considered non-academics’ trust in sociologists, legitimacy has become entwined with the idea of a value-free, ‘objective’ sociology. However, broader philosophical/sociological work suggests that credibility signals are more complex, with, for example, non-partisanship being separate to epistemic responsibility. In this article, I explore the nature of ‘credibility work’ in practice via interviews with 15 prominent English sociologists, making three contributions. First, I find that some sociologists deliberately pursue credibility, a phenomenon largely ignored in previous research. They do this primarily by ‘performing’ non-partisanship or epistemic responsibility within interactions. Second, this credibility work does not require the pursuit of ‘objectivity’; sociologists can signal epistemic responsibility despite partisanship, or pursue ‘dispassionate advocacy’. Third, the extent and nature of credibility work varies by context; indeed some sociologists benefit from partisanship, while others feel no need for credibility work. I conclude by stressing the need for further research.

Introduction

Sociologists have said surprisingly little about other people’s trust in them, and how this may be affected by their own behaviour. This is an unfortunate lacuna given mounting public concern about ‘post-truth’ politics. Even if the term ‘post-truth’ is misleading (Jasanoff and Simmet, 2017), there is evidence that trust in scientific experts has declined (at least among US conservatives; Gauchat, 2012), and there is an increasingly prominent populist politics that brackets (social) scientists in with a malign ‘elite’ (Inglehart and Norris, 2016). Moreover, even before Trump, Brexit, Orbán et al., considerable academic attention was devoted to trust in science (Gauchat, 2012; Gieryn, 1999; Shapin, 1994; Silke, 2012) – yet such concerns were rarely taken up by sociologists.

This is not to suggest that trust in sociology has been entirely ignored by sociologists. Our discipline has always been subject to debates about whether we should present ourselves as a ‘dispassionate science’, and while these debates are wide-ranging, legitimacy does feature within them (see below). However, these debates have two limitations. First, legitimacy is almost exclusively mentioned by those pursuing a value-free, ‘objective’ sociology. This vision is rejected on epistemological and moral grounds by those pursuing a more value-committed sociology – but in the process, they ignore issues of legitimacy. As a result, concerns about trust become conflated with scientism. Second, there is no empirical attention given to what sociologists actually do. Indeed, even beyond sociology, there are few if any studies that consider whether (social) scientists deliberately pursue credibility beyond the academe.

In this article, I empirically examine whether prominent English sociologists engage in such ‘credibility work’, based on interviews with 15 high-impact sociologists. This enables three contributions. First, I find that some sociologists deliberately pursue credibility, a phenomenon largely ignored in previous research. They do this primarily by ‘performing’ non-partisanship or epistemic responsibility within interactions, rather than having a priori strategies. Second, this credibility work does not require the pursuit of ‘objectivity’. Instead, sociologists can advocate for policies or use vivid language while presenting themselves as dispassionate; or present themselves as partisans with epistemic responsibility. Third, the extent and nature of credibility work varies by context; indeed some sociologists benefit from partisanship, while others feel no need for credibility work. Before this, however, I first justify my claim that sociological debates tend to conflate concerns about trust with scientism.

Debating the Legitimacy of Sociology

My argument here is that sociologists only consider trust in sociology when debating the nature of the discipline, and even here debates about trust are partial. To provide evidence for this, we can examine the most significant recent Anglo-American debate: Burawoy’s ‘public sociology’. This was a call for ‘sociology that seeks to bring sociology to publics beyond the academy, promoting dialogue about issues that affect the fate of society’ (Burawoy et al., 2004: 104), which numerous sociologists responded to across special issues (Social Forces 82(4), Social Problems 51(1), the British Journal of Sociology 56(3), Critical Sociology 31(3) and the American Sociologist 36(3–4)) and edited collections (e.g. Clawson et al., 2007; Jeffries, 2009). What concerns us here, though, is the way in which trust in sociology is invoked.

Trust is mentioned by several critics of public sociology. Some critics focus on the inherent trustworthiness of sociological accounts, which they argue become untrustworthy if we are too close to civil society (Goldberg and Van den Berg, 2009; Hammersley, 2017; Nielsen, 2004; Tittle, 2004). But more importantly here, critics claim that even robust public sociology will be credibility-damaging because public sociologists display left-wing values that may not be shared by their audiences (Goldberg and Van den Berg, 2009; Tittle, 2004; Turner, 2005), and because they blur the distinction between knowledge and advocacy (Goldberg and Van den Berg, 2009; Hammersley, 2017; Tittle, 2004). As Turner (2005: 30) puts it:

if sociologists simply throw their ideological hats into the ring, spouting off their own moral judgments, their credibility will be lost; and political counterattack will be easy. We become just another set of talking heads engaged in a moral crusade.

In defence, Burawoy accepts that trust in public sociology is helped by its coexistence with ‘professional sociology’, which ‘is not the enemy of . . . public sociology but the sine qua non of [its] existence – providing both legitimacy and expertise’ (Burawoy, 2005a: 10). However, he otherwise ignores legitimacy, instead counter-attacking on other grounds:

How often have I heard professional sociologists dismiss public sociology as a euphemism for partisan sociology – and how public sociology threatens sociology’s legitimacy as a science . . . There is nothing wrong with activism, they say, but it should have nothing to do with sociology which, implicitly or explicitly is assumed to be value neutral. Again the tactic is to pathologize the enemy while simultaneously idealizing the self. (Burawoy, 2005b: 424)

In other words, legitimacy is only mentioned by those pursuing value-freedom, and is ignored by those who reject it. A similar pattern can be seen in pre-1920s US sociology, split between those attempting to reform society and those pursuing knowledge, as Burawoy notes (2005a). This debate was set against a background of legitimacy challenges, in which some radical sociologists were hounded out of the discipline completely (Furner, 1975; Smith, 1994). Just as in recent debates, legitimacy issues were only raised by those pursuing a more dispassionate sociology. For example, Du Bois argued that ‘the frequent alliance of sociological research with various panaceas and particular schemes of reform, has resulted in closely connecting social investigation with a good deal of groundless assumption and humbug in the popular mind’ (albeit before largely changing his mind; see Bright, 2018: 2232).

My point here is not to judge the strength of these arguments per se, but rather to show how concerns for legitimacy have become entwined with scientism. While this scientism has been robustly countered, this has been on the grounds of philosophical coherence and moral responsibility rather than legitimacy. Epistemologically, sociologists have written extensively about how to go beyond a binary choice between scientism and partisanship via, for example, Bourdieusian accounts that produce robust knowledge that reflect upon the categories they use to study the social world. And when it comes to trust/legitimacy in practice, sociologists similarly go beyond such a binary choice, as I will show. However, within the academic record, our discipline’s discussion of trust remains restricted to calls for a scientistic sociology, as part of a contrast with value-laden advocacy.

Further Accounts of Trust in (Social) Science

While sociologists have barely considered these issues, we can learn from wider accounts of trust in science/expertise, particularly Science and Technology Studies (STS), for which credibility is a ‘fundamental topic’ (Shapin, 1995: 257). This analyses scientific disputes without any reference to their truth/falsity; we cannot say that the accounts of particular scientists win out because they are true, instead investigating why people regarded them as true; after which ‘the study of credibility then became simply coextensive with the study of knowledge’ (Shapin, 1995: 256). A particular concern in STS has been the way in which some individuals gain credibility through demarcating themselves from non-scientists (often invoking Mertonian norms such as disinterestedness and organised scepticism), known as ‘boundary-work’ (Gieryn, 1983). This is clearly relevant to sociology, and Gieryn himself (1999) has documented the efforts that social scientists made in the 1960s to persuade the US Congress that their work counts as ‘science’.

However, there are good philosophical reasons not to trust all scientists solely by virtue of their scientific status (Anderson, 2012; Goldman, 2001), and we see in practice that the credibility of individual scientists is often contested (of many examples within STS, see, for example, Campbell, 2004; Hilgartner, 2000; Shapin, 1994; Silke, 2012). This sometimes follows the binary distinction between advocacy and objectivity: studies of policymakers in real-world settings show they prefer what they consider to be objective evidence over ideologically driven, biased advice (Haynes et al., 2012; Jacobson and Goering, 2006; Weiss, 1980). Similarly, experimental studies among the general public show that if scientists make recommendations, they are viewed as less impartial (Beall et al., 2017; Elliott et al., 2017; Kotcher et al., 2017).

However, to fully understand the credibility of social scientists, we must go beyond a binary model. Philosophers of science and empirical STS studies both suggest that there are multiple dimensions to the trustworthiness of scientists, of which two are particularly important:

• Motivation: there is wide agreement that someone who has a vested interest in a particular result should reasonably be viewed as less trustworthy (Anderson, 2012; Goldman, 2001; Shapin, 1994: 223).

• Epistemic responsibility: credibility also depends on what Elizabeth Anderson (2012: 146) terms ‘epistemic responsibility’: ‘whether testifiers are responsive to evidence, reasoning, and arguments others raise against their beliefs . . . The mark of epistemic responsibility is responsive accountability to the community of inquirers.’ Others have similarly emphasised the role of visible signs of being ‘responsive to evidence’ (Douglas, 2015; Goldman, 2001: 94–97; Longino, 2002: 130–133; Shapin, 1994: 223).

This raises the possibility of sociologists deliberately pursuing credibility, but not through the naive pursuit of ‘objectivity’. For example, they could admit that they have a stake in the result, while simultaneously signalling that their conclusions are firmly grounded in the evidence rather than in wishful thinking. The binary conception of credibility that we see in sociological debates (above) is therefore inadequate.

Yet if we want to understand sociologists’ deliberate pursuit of credibility with different non-academic audiences – ‘credibility work’ – these literatures are still unsatisfactory. First, STS focuses almost exclusively on the natural sciences; we would expect credibility debates in the social sciences to operate differently, yet social sciences are studied by only a small and marginal part of STS (Pereira, 2019). Second, there is almost no research investigating whether credibility is consciously pursued, even by natural scientists. Three studies (one on climate science, two on public health) have suggested researchers are aware of credibility, and felt that they needed to present themselves as objective (Gundersen, 2018; Haynes et al., 2011; Jacobson and Goering, 2006). However, the only study that includes social scientists is the study of UK health inequalities researchers by Katherine Smith (2010, 2012).

These few studies do hint that ‘credibility work’ exists, and that it goes beyond a binary advocacy-objectivity model. Smith describes a divide between ‘critical’ and ‘policy-relevant’ researchers, but emphasises that some critical researchers consciously blunted their critiques to some audiences in order to maintain ‘some level of policy credibility’ (Smith, 2012: 157). And Haynes et al. (2011: 1050) argue that the debate about advocacy vs. objectivity is in practice often not about advocacy per se, but a ‘more subtle distinction between data-informed, moderated commentary and ideologically-driven proselytising which can ‘do more harm than good’. Yet a detailed study of whether prominent social scientists consciously pursue credibility – and if so, the nature of this work – is lacking. In this study I fill this gap.

Methods

Design

My aim here was to explore whether such ‘credibility work’ existed among English sociologists, and if so, the extent to which sociologists in practice adopt more nuanced positions than the extremes of ‘advocacy’ vs. ‘objectivity’. The research design was based on semi-structured interviews, which provide the space for respondents to describe credibility work in their own words, as well as providing the opportunity to understand it within the context of their personal biographies. By design, interviews enable us to investigate the conscious pursuit of credibility; in the Discussion below I consider how to interpret this within the broader context of behaviours and attributes that may influence audiences’ perceptions of credibility.

My sampling approach used the logic of a ‘most likely’ case study design (on using case study logics for interview sampling, see Small (2009), although the design can easily be re-described as purposive sampling). That is, for this small-scale exploratory study, I aimed to select ‘information rich’ (Patton, 2002: 242) cases in which credibility work was particularly likely. This was for two reasons: first, because I aimed to discover whether conscious ‘credibility work’ exists at all; and second, because I aimed to look at the nature of credibility work, for which evidence was only available from cases in which credibility work was present. In practice, I used policy/public impact as an outwardly observable sign of cases in which credibility work was likely. This was done via searches of impact case studies from the Research Excellence Framework (REF) 2014, searches for mentions of sociologists in online news media and examining prominent policy advisory roles.

There are limitations to this sampling approach. Having selected high-impact sociologists for whom credibility work is more likely, I cannot talk about the extent of credibility work among sociologists in general. Moreover, given the widely noted problems of the REF impact agenda, this risks a sample that is relatively willing to make accommodations to policy elites, neglecting the more radical end of sociology that works outside of these elites towards longer-term radical change (Smith and Stewart, 2016). It may well be the case that credibility work outside of this sample takes a different form; as we will see below, credibility work depends, inter alia, on the preferences of each individual sociologist and on the audience judging their credibility, and I make no claim to have captured the full range of sociologists or audiences here. Nevertheless, the design is appropriate given the scale, exploratory nature and aims of the research, and in the Discussion below I consider fruitful avenues for future research.

I did however vary the sample according to other dimensions that are likely to influence credibility work, namely methodology and field of study. My expectation – borne out below – was that different fields of study and different types of evidence create different pressures and opportunities for credibility work, and capturing this variability within the sample enables an exploratory analysis of the role of sociological context in credibility work (see below). Reflecting this, I focused on those studying sociological terrain from any disciplinary base, including criminologists, social policy scholars and social geographers. Ultimately the interviewees covered a variety of sociological topics, including inequality; the welfare state; ethnicity/integration; addictions; disability; family policy and social work; and charitable giving.

The final sample included 14 interviews with 15 sociologists (two long-term collaborators being interviewed together). Interviewees include some of the most well-known and influential sociologists in the country; most (13) were Professors, and the majority (12) were involved in REF impact case studies in 2014. Eight of those invited to participate declined, five of whom were female, leaving only three female participants; I discuss this when considering epistemic injustice below.

Fieldwork

Interviews were predominantly conducted face-to-face at a location of interviewees’ choice (one was conducted by video-call) in August–September 2018, and lasted 50–102 minutes (averaging 73 minutes). Interviews were transcribed and then analysed through a mixture of thematic coding and case summaries. Thematic coding in NVivo enabled me to organise participants’ accounts around my research questions, allowing differences between accounts on specific issues to be explored in-depth. Complementing this, case summaries allowed accounts on specific issues to be placed in the context of each interviewee’s account/biography, avoiding the fracturing tendency in purely thematic analyses. Interviewees were given a choice about the level of anonymity they would prefer; interviewees below are referred to by name where they gave consent and where the individual’s biography is helpful for understanding their responses. Quotes either give life to typical responses or highlight unusual cases, while trying to use the full set of accounts. I return to limitations and the need for further research in the Discussion below.

Half of the interviewees knew me personally, which helped recruitment and rapport and allowed greater space for interviewer and interviewee to challenge one another. This does however raise additional ethical issues around consent and anonymity; additional safeguards were provided, and the project was approved by the University of Kent research ethics committee (SRCEA id 209). So-called ‘elite interviews’ do not raise fundamentally different issues to other interviews, but often have unusual power dynamics (Smith, 2006), and there is likely to be a greater-than-usual element of conscious self-presentation here, as I return to below.

In the interviews, I directly asked respondents about whether they took deliberate steps to be seen as trustworthy, as well as asking how they thought others perceived them, any challenges to their credibility and whether they gave policy advice. (These were adapted to the particular biography of each interviewee, based on prior desk research.) This crucially allowed me to probe the absence as well as the presence of conscious credibility work, rather than leaving silence as ambiguous. While such directional questioning could be leading in other contexts, this was suitable for this group of prominent academics, who were prepared to challenge my views – as I show below – and who expected to be allowed to engage directly with the project’s hypotheses.

Results I: The Nature of Credibility Work

There were plentiful examples of respondents consciously presenting themselves as credible to non-academic audiences – ‘credibility work’. To understand this, we must examine how different contexts provide more/less pressures for credibility work, which is covered below. First, though, I examine the two main types of credibility work that participants described: non-partisanship and epistemic responsibility. (We return below to other, rarer forms of credibility work.)

Non-Partisanship

There was clear evidence that some high-profile sociologists consciously tried to present themselves as ‘non-partisan’ – that is, they suppressed signalling their personal value positions to avoid being seen as a vested interest. This could take various forms, whether avoiding advocacy (e.g. not making direct recommendations in research reports; avoiding round-robin letters to newspapers), not joining a political party or not identifying closely with social movements. It could also be performed within social interactions:

When I was interviewed for [an official role], I made a conscious effort to present myself as quite a ‘small c’ conservative person, because I knew they probably wouldn’t select me if they thought of me as being too radical. So I went in the soberest [outfit] I’ve got, I presented myself in the most scientifically respectable way possible . . . It wasn’t about the credibility of my science, it’s about who you are and what sort of person people perceive you to be. And you can deliberately create that impression.

At times, such credibility work could be consciously insincere, with avowed radicals adopting a neutral’s clothing to avoid being dismissed – ‘almost pretending you’re not an old leftie’, as one put it. But for most, conscious displays of non-partisanship were a way of demonstrating independence of thought rather than pure neutrality. Paul Gregg – an adviser within New Labour – explained that he

would always wear my most ripped pair of jeans whenever I met Tony Blair [then-Prime Minister] . . . That [outfit] created the distance. I wasn’t one of their suits . . . I wanted to create that kind of image that I’m semi-independent, I can be trusted but I’m a voice that’s trying to be . . . distant and sometimes provocative.

Epistemic Responsibility

The other common dimension of credibility work was performing ‘epistemic responsibility’ – being ‘responsive to evidence, reasoning, and arguments others raise against their beliefs’ (Anderson, 2012: 146) – which could be signalled in various ways. First, nearly every respondent emphasised that they were based on the evidence, often contrasting this with the ‘rent-a-quote’: ‘anyone can gob off and then you go on Newsnight and say something provocative and you’re a star, but that’s not what I do’. Second, several visible partisans felt that engaging with political opponents in a respectful matter – particularly genuine listening and engagement – was a powerful way of signalling trustworthiness. Third, most tried to manage their emotional tone, particularly anger. Anger was described as ‘going off on one’, ‘ranting’ or ‘losing it’; ‘it would make great telly but it won’t get the message across, they’ll just think “look at that fool”’. Avoiding anger was sometimes felt to be a particular challenge on Twitter, which several respondents had therefore stopped using after potentially credibility-damaging failures to manage their emotional tone.

Finally, epistemic responsibility could also be performed through the very argument that was being presented. The social scientists I spoke to often ‘tried to be as fair as possible’, to show that ‘you will change your mind when the evidence changes’ and to ‘head off’ any damaging criticisms that were likely to be raised; others visibly considered uncertainties and alternative interpretations. These were partly presented as good academic practice (see below), but also as deliberate signals of credibility. Hence for one academic who was ‘seen as biased’ from previous critical work, it was important that they ‘bent over the other way’ in trying to consider whether their new qualitative research could possibly support the Government’s claims. John Hills – a widely trusted LSE professor working on inequality – described the challenge of considering alternatives in short media appearances: ‘you are trying to get over the idea that although you’ve only got two minutes to talk, you have thought about the arguments going in different directions’.

Dispassionate Advocacy

Yet belying the binary opposition of ‘advocacy’ vs. ‘objectivity’, epistemic responsibility could be performed while social scientists advocated for particular policies, and/or used emotional, vivid language. With respect to emotional tone, for example, there was rarely a complete suppression of ‘emotion’ in order to emphasise ‘reason’. Emotions were often seen as appropriate: ‘you don’t burst into tears, but you make it quite clear that this is emotionally important, it’s not just an interesting intellectual problem how many children die’. But more than this, as long as anger itself was avoided, then emotional engagement could be valuable in putting together a persuasive ‘call-to-arms’: as one put it, ‘to some extent you have to back off the distal or objective academic and be a little bit more out there and in people’s faces’. The ideal position was sometimes to be ‘passionate’ but not angry, or in the evocative phrase that Alex Stevens used, to express ‘cold fury’ – an authentic and appropriate display of emotion that nonetheless signals that values are being managed.

What is perhaps most striking is how independent these considerations were from ‘advocacy’ in the narrow sense of arguing for a particular policy option. Most respondents were willing to recommend particular courses of action, even while simultaneously trying to position themselves as non-partisan and signal epistemic responsibility. In every case, they did this by emphasising how their advocacy was evidence-based; for example, one argued that their recommendations were ‘not about a normative position’, and instead they were ‘trying to let the evidence speak’. We can perhaps term this ‘dispassionate advocacy’, noting the clear affinities with the UK’s ‘evidence-based policy’ movement.

Results II: Clarifying Credibility Work

Before proceeding further, we need to clarify that ‘credibility work’ is an ‘analyst’s category’ (Collins, 2008) – it is built upon sociologists’ accounts of their behaviour and its motivation, but it is not a term that they themselves use. This is important in two ways. First, the pursuit of credibility was not a dominant aspect of these sociologists’ daily lives. It was not uncommon for academics to explicitly reject the idea that they ‘went around thinking “I’ve got to do this in order to gain more credibility”’, or that they did what one dismissively called ‘strategic credibility management in the way a PR agency might’. Nevertheless, there were countless examples of small, occasional, ad hoc but deliberate decisions in the light of a risk of being discredited, as one interviewee put unusually clearly:

I don’t think I’ve gone through with a headline strategy of saying, ‘I need to build up my credibility as an expert’ . . . It’s more like, you do make these micro-decisions on a day-to-day basis as things arise that could affect your credibility, and your guiding principle is ‘I need to maintain my independence and credibility’.

Credibility work was therefore largely a matter of performance – one participant even described making ‘a play out of being credible, and in very practical performative terms’. The metaphor of performance is often associated with Goffman’s dramaturgical metaphor, and has been used to capture the way that professionals – including scientific advisers (Hilgartner, 2000) – present themselves. However, I here more narrowly refer to ‘performance’ to capture the ways in which credibility was consciously signalled in the midst of interactions with non-academics, which better captures most credibility work than the idea of credibility work as a priori strategic thinking.

A second clarification is that occasional interviewees directly resisted the concept of credibility work: ‘Isn’t that interesting, you’re talking about the presentation of self here aren’t you? Well to be honest I’ve never felt that. I don’t think I’ve got to prove myself.’ Yet such statements should not be taken at face value; the same respondent shortly afterwards described consciously considering credibility (‘I’ve tried, personally and my teams and the places I worked, to make sure we don’t get dismissed’). Instead, this shows how admitting to credibility work could itself be discrediting; as another said, ‘playing to your audience in a small way is okay, but if you do it in a big way you become untrustworthy, cause you’re just conning them’. As this respondent stressed, the impression of credibility overlapped with behaving with integrity:

The primary route for gaining trust is to be trustworthy. And so you don’t bullshit, you don’t fiddle your figures, you don’t jump over challenging bits in your argument, you don’t overstate your conclusions, and then you make it more likely that people will trust you, cause you earn it.

This concern with sincerity resonates with a common view in STS that scientists are ‘actors [who] use stage management techniques’ to present themselves as disinterested, and that these ‘prevent their audiences from seeing this performance as the authoritarian ruse that it is’ (Campbell, 2004: 433) – which mirrors a common critique of Goffman-esque ‘impression management’. There were rare examples here where a concern for others’ perceptions meant that a performance was insincere. However, there is nothing in Goffman’s work that requires performance to be insincere (Tseëlon, 1992), and in practice, I found that credibility work was overwhelmingly sincere. Moreover, it has been influentially argued by Naomi Scheman (2001: 35) that scientists have an obligation to communicate their trustworthiness in order to justify their societal role. There is nothing in my concept of ‘credibility work’ that implies insincerity, and I share Scheman’s view that credible sociologists have an obligation to demonstrate their credibility.

To be clear: ‘credibility work’ does not incorporate behaviours that are not motivated by the pursuit of credibility, even if they are credibility-enhancing (e.g. good methodological practice). But we must accept the possibility that some sociologists underplayed the extent to which behaviours were motivated by credibility, given that some associated credibility work with insincerity.

Results III: Credibility Work in Context

Trustworthiness is not a unidimensional phenomenon, instead depending on both the audience and the type of trust to be given (Haynes et al., 2012: 6; Jones, 2012). The non-academic audiences considered here include, inter alia, politicians, civil servants, civil society groups, the media and the public (none of whom are homogeneous), which means that the pressure for credibility work per se and the way that credibility is judged will vary. In this final section I therefore turn to how the context of credibility work affects its nature and extent.

Where Partisanship Connotes Credibility

We have seen that some sociologists deliberately signalled non-partisanship, but other sociologists spoke of situations in which partisanship helped their credibility. This was particularly the case when being drawn into the inner circle of powerful politicians (in all cases here, Labour politicians), who were concerned with ‘political credibility’ – whether the sociologist was ‘on their side’ or ‘ideologically trustworthy’:

Particularly within the Labour Party, it is ‘are they one of us?’ syndrome, and at some point what they want from you is to know that you’re going to come out with a range of ideas that are within their blanket brackets of the acceptable, or the politically sympathetic.

This was changeable; insiders could be cast out if they publicly backed a policy that was outside political acceptability, or if the dominant ideological strand of the Labour Party itself changed (senior advisers in the Blair/Brown government lost their political credibility under the Corbyn leadership). But what matters for our purposes is that partisanship could be an asset rather than a liability.

Partisanship could also help gain trust in other ways. Both politicians and the media would sometimes search out expert advocates of a particular position, wanting them to make their case as persuasively as possible. And while the sample design privileged sociologists who worked with politicians/policymakers/the media rather than smaller advocacy groups (see above), there were nevertheless examples of partisanship helping via a chain of credibility:

I think there was a time . . . where really certain researchers were seen very much as allies of advocates. And then because we were also working in a European space, those civil society advocates were also very prominently placed in policy debates. Always in the room. And so their endorsement of us as allies would have perhaps made a difference.

However, partisanship rarely constituted ‘credibility work’. In only one case did a sociologist deliberately tried to garner credibility via partisanship (this was a respondent on the right of the Labour Party who made a point of signalling their longstanding party membership). Moreover, partisanship was largely helpful for political credibility – being trusted to frame problems and solutions in politically useful ways – rather than the epistemic credibility that concerns us here. This does not mean that epistemic credibility is irrelevant to sociologists working in partisan contexts. I have already given the example of Paul Gregg talking to New Labour politicians (who signalled that ‘I can be trusted but I’m a voice that’s trying to be . . . distant and sometimes provocative’), which others echoed: ‘Talking to politicians, you’d get tested out on two levels . . . [One is], “this is somebody who knows about his subject” . . . The other one is “can I trust this guy politically?”’

The role of partisanship in credibility is therefore complex, and I return below to how future research could explore this further. But from the interviews here, it is clear that sociologists engage with different non-academic audiences in different roles, which change the role of partisanship. Some sociologists sometimes deliberately pursued epistemic credibility by signalling non-partisanship. But at other times – particularly when becoming partly political actors – partisanship was valuable because it increased sociologists’ political credibility, even if this occurred alongside other credibility work to maintain their epistemic credibility.

The Pressure for Credibility Work

Context affected not just the value of partisanship, but the very pressure for credibility work itself. I consider this through three lenses: (1) where credibility work was unnecessary; (2) where credibility was assumed rather than achieved; and (3) where pressures for credibility work were ignored.

Where Credibility Work Was Unnecessary

Some sociologists faced contexts where their credibility was vigorously challenged, but others faced less challenge, and in two cases felt no challenge whatsoever. While these two respondents varied in many ways – for example, seniority, institutional prestige – they were united by working within fields that had large and powerful practitioner communities, which they saw themselves as partly representing, and which in return supported their credibility. As one put it:

I didn’t want to go in and be authoritative in telling the sector what to do. I wanted to be authoritative in articulating the sector’s views, adding my own intellectual value, whatever it was, and being authoritative to the politicians.

They were also similar in that they made policy recommendations that were not party-political. It is not that they hid their (left-wing) political positions, but rather that their ideas were widely accepted by practitioners, and had value to politicians across the political spectrum. One actively tried to present themselves to politicians as ‘not politically partisan but aligned with their objectives: so here’s how I can help you do what you want to do’. The other had taken a role for the (Conservative-led) Coalition Government, and reflected that ‘to some degree the Government were happy to accept my recommendations because they saw them as fitting within [their agenda]. So they liked them for the wrong reasons.’

In these contexts, credibility was the default; one simply said, ‘I don’t feel a huge burden of proving my objectivity.’ While this extreme was rare among the sociologists I spoke to, there were elements reported by others; for example, finding less need for credibility work where ideas were presented as apolitical, or doing less credibility work where fields became less politically contentious over time. To understand the (non-)existence of credibility work, we must understand the politicisation of a given sociologist’s field, which determines the reputational pressures they are likely to face.

Assumed Credibility

Aside from their deliberate credibility work, some sociologists seemed to benefit from audiences’ a priori assumption that they were trustworthy. This partly reflected their research methods, with several qualitative researchers feeling that their research was under-valued: ‘qualitative work is easily dismissed because it’s not [whispered] “science”’. One sociologist even resorted to doing qualitative research with quantitative sample sizes and then quantitising their results to maximise their credibility in political debate: ‘I’ve got little graphs that show [a key finding] that we put in the final report because they want to see this, even though it’s not really the heart of what we’re doing.’

It is nonetheless over-simplistic to treat this as a quantitative/qualitative divide. A sociologist doing more complex quantitative modelling pointed out the credibility challenges that complexity introduced, reflecting that ‘it’s perhaps easier for people who are doing a simpler piece of work’. In contrast, those doing simple quantitative research could choose to provide (what was perceived to be) unmediated access to descriptive quantitative data. Danny Dorling – a high-profile academic working on inequalities – argued that the ‘best evidence for [my credibility] strategising’ was that he put all of his underlying data on the internet from the mid-2000s; he retrospectively felt that ‘something must have happened that made me think, “You can easily be attacked on this.”’ This is a different type of credibility work to the previous section: it is not that Dorling was here demonstrating his trustworthiness, but rather that he attempted to reduce the very need for trust (as he said when we discussed one claim, ‘you can look at the data and say I’m wrong’).

Credibility could also be assumed based on personal attributes. Foremost among these were markers of academic prestige such as the title ‘professor’ or being attached to a prestigious university; as the Oxford-based Danny Dorling put it, ‘for the very right wing of the British press, Oxford can do no wrong. It’s where they went and it produced them, it made them, so it must be great.’ Sometimes sociologists deliberately publicised such markers in order to enhance credibility, although overall this was a minor element of ‘credibility work’. I also expected disadvantaged groups – for example, women, ethnic minorities – to suffer ‘epistemic injustices’ (Pereira, 2019) that heightened the importance of credibility work, but most respondents explicitly rejected this. This may partly be explained by the aforementioned under-representation of women in the sample; however, those women I interviewed did not feel that their gender devalued their credibility, with epistemic injustice only being raised by two male respondents in the context of unusually patriarchal settings such as the House of Lords.

However, we must be tentative about these claims, as I did not speak to sociological audiences, and sociologists themselves may be an unreliable guide to audiences’ reactions, or unaware of the reasons that their strategies were successful. For example, it is clearly not the case that any audiences really have ‘unmediated’ access to Danny Dorling’s data, given his role in framing, designing, collecting, cleaning and interpreting his data. Given that other publicly available data are contested (e.g. official labour force/immigration statistics), it is unclear – both to myself and probably to Dorling himself – what conditions are necessary for his strategy to be successful. Moreover, given that I selected sociologists on the grounds of high impact, the female sociologists I spoke to may be unrepresentative of the experiences of other female sociologists. I return below to how future research could further explore this ‘assumed credibility’.

Ignoring Pressures for Credibility Work

Finally, it should go without saying that sociologists are active agents who can choose how to respond to the pressures they face. Several interviewees described doing less credibility work as their careers progressed, partly because they became sceptical about the power of credibility work:

I feel much more willing to pin my political colours to the mast than I used to. I used to feel it was quite useful to be able to say, ‘I don’t belong to any political party, here’s my evidence, make of it what you will’, almost like pretending you’re not an old leftie. But I don’t feel that anymore . . . [With emphasis] See the thing is I don’t think really that people change their minds that much.

Given that I deliberately chose a sample in which credibility work was likely, it may well be the case that sociologists outside of my sample are even more likely to ignore the pressures for credibility work.

Discussion

Sociologists have rarely considered how other people’s trust in them may be influenced by their own behaviour; to the extent they have, it is considered only by those arguing for a value-free, ‘objective’ sociology, which many sociologists reject on epistemological or moral grounds. Moreover, despite the wealth of insights from STS and philosophy of science, almost no previous studies have examined whether social scientists consciously try to shape their credibility to non-academic audiences (‘credibility work’). The present article therefore reports findings from an exploratory study of whether sociologists conduct such credibility work, and if so, what form this takes, using qualitative interviews with 15 prominent English sociologists. The findings make three contributions to the literature.

First, I found clear evidence that credibility work exists: some sociologists do consciously pursue credibility with non-academic audiences, mostly by performing credibility via small, ad hoc decisions to deflect credibility threats. This is itself a contribution: even for natural scientists, empirical research has largely ignored the existence of credibility work (the only exceptions being Gundersen, 2018; Haynes et al., 2011; Jacobson and Goering, 2006), and only Smith (2010, 2012) examines this – briefly – for social scientists. I found that this credibility work can take various forms, including signalling partisanship, increasing transparency, signalling markers of prestige or quantitising qualitative data. Predominantly, though, credibility work took the form of signalling non-partisanship or epistemic responsibility (that is, being visibly responsive to evidence and argument; Anderson, 2012).

Second, credibility work goes beyond the way that sociologists discuss trust within academic debates. In these debates, trust/legitimacy is only considered by those pursuing scientism, who argue trust requires a value-free, ‘objective’ sociology, rather than partisan, untrusted advocacy. In contrast, sociologists in practice adopted more nuanced positions; for example, being partisan while signalling epistemic responsibility. Many also pursued ‘dispassionate advocacy’: they recommended particular courses of action but nevertheless tried to avoid being labelled as a pure advocate. They did this by stressing the connection between evidence and their recommendations, emphasising their non-partisan status, engaging with those with different values, demonstrating their willingness to consider alternative accounts or managing their emotional tone.

Finally, the extent and nature of credibility work varied. While some sociologists signalled non-partisanship to gain credibility, others were aware that partisanship helped gain political trust. While most sociologists felt a pressure to deflect credibility threats, a few felt less/no pressure, particularly when making apolitical recommendations supported by practitioners, or where they benefited from an a priori assumption of credibility. And even if sociologists were aware of credibility threats, they could choose to ignore them. We should therefore expect credibility work to be heterogeneous across contexts, and it is perhaps unsurprising that the rare quantitative studies of the effect of different types of credibility work find mixed results (e.g. for the effect of revealing scientists’ values; Beall et al., 2017; Elliott et al., 2017; Kotcher et al., 2017).

Limitations

This was an exploratory study, and there is a need for future research that goes further in both sample and method. Regarding the sample, I spoke only to English sociologists who were high-impact in a particular sense (those with REF impact submissions/prominent advisory roles/regular media appearances). This risks a sample that is relatively willing to make accommodations to policy elites (Smith and Stewart, 2016); it is likely that other sociologists engage differently with non-academic audiences, using theoretical rather than empirical insights to work with grassroots organisations towards long-term radical change. Future research should therefore include a broader and more international sample to see if credibility work exists in other contexts, and if so, whether it takes a different form to that found here.

Regarding method, I chose interviews so that I could explore whether sociologists were motivated by credibility, but other methods are now needed. First, future studies should complement interviews with detailed observational studies of how sociologists perform credibility in different contexts (ranging from face-to-face meetings to social media). Second, credibility work may be unsuccessful, with credibility instead depending more on other behaviours/characteristics – making it crucial to study audience reactions. Third, it seems likely that sociologists are selected for being credible, which reduces the need to consciously pursue credibility. This may explain why for example prominent female academics did not report gender-related epistemic injustice (as their participation in the research was conditional on having overcome this). It would therefore be valuable to conduct a longitudinal project based on a heterogeneous cohort of early career social scientists, which examined both their characteristics and practices over a sustained period of time, and the ways in which different audiences responded to them.

This was an exploratory study for an issue that is curiously under-studied by sociologists. It is already clear that sociologists have more subtle choices available to them epistemologically than ‘advocacy’ versus ‘objectivity’; in this article, I hope to have demonstrated that the same is true when considering credibility, even if the impacts of different practices (and the trade-offs inherent to them) are questions left to future research.