Think Tanks, Television News and Impartiality

Is the use of think tanks ideologically balanced in BBC news and current affairs programming? This study answers this question empirically by establishing which think tanks are referenced in different BBC programming in 2009 and 2015, and then classifying them according to their ideological aims (either left, right, centrist or non-partisan). We draw on a sample size of over 30,000 BBC news and current affairs programmes in 2009 and 2015 to measure how often these think tanks were mentioned or quoted. Overall, BBC news reveals a clear preference for non-partisan or centrist think tanks. However, when the Labour Party was in power in 2009, left and right-leaning think tanks received similar levels of coverage, but in 2015, when the Conservative Party was in government, right-leaning think tanks outnumbered left-leaning think tanks by around two to one. Overall, our findings add weight to a pattern emerging from a number of recent academic studies that show, despite its undoubted commitment to impartiality, BBC news programming has shifted its centre of gravity to the right. We argue that broadcasters need to be more independently aware of how stories emerge, and how issues and sources should be balanced and explained in an increasingly partisan news environment.

In an era of "fake news" and "post-truth politics", one of the key challenges facing broadcasters is maintaining these core values while being part of a changing news ecology that includes print, online and social media, where there are no obligations or traditions of fairness or impartiality. This has put notions of impartiality under increasing stress, in three ways.
First, pressure on broadcasters from other parts of the news ecology has amplified, with politicians and partisan media-notably newspapers who resent competition from a public service news provider-becoming increasingly aggressive in attacking the BBC for alleged bias (Lewis 2015). Much of this pressure is directed at the BBC, partly because it is the most popular source of news in the United Kingdom, and partly because its public funding makes it politically vulnerable.
The most visible source of pressure comes from politicians and newspapers on the political right-many of whom are unsympathetic to the principle of public service broadcasting (Lewis 2015). In this politically charged climate, the power of the complainant tends to matter more than the quality of the evidence. Carefully argued, evidence-based and peer-reviewed academic research is easier to ignore than a diatribe by a government minister or the editor of the Daily Mail. The danger here is palpable-if broadcasters are to appease their more vocal critics, they will be more sensitive to accusations that they lean to the left and less concerned about leaning to the right.
Second, evidence suggests that, regardless of political pressure, a partisan UK press plays an agenda-setting role for broadcasters (Cushion et al. 2016a). The ecology of news in the United Kingdom means that broadcasters see the press as a key source of news stories, and a bellwether of public debate. But because most UK newspapers favour parties and report issues in ways that favour right-wing perspectives (many of them-the Mail, Sun, Express and Telegraph-vociferously so), this runs the risk of broadcasters assuming a centre of gravity that reflects the partisan press landscape (Renton and Schlosberg 2017). Research on the coverage of the 2015 General Election campaign, for example, suggested that broadcasters may have been pushed towards a news agenda that favoured the Conservative Party (Cushion et al. 2016b).
Third, because broadcast journalists are part of a larger news system, they are inevitably influenced by the more partisan sector, which, while it may not share values of impartiality (or, in some cases, accuracy), often has a similar set of journalistic assumptions about what makes a good story. Retaining impartiality amongst the maelstrom of claims, public relations and political vested interests requires a sense of detachment that is difficult to maintain (Lewis, Williams, and Franklin 2008). When President Trump claimed on Twitter that his predecessor-Barack Obama-had wiretapped him, for example, it was widely reported (such as a Fox News banner reading: "WIRETAPPING FROM THE WH"). Simply by reporting the story, a BBC breaking news tweet went some way in legitimating it: "President Trump urges Congress to examine whether Obama abused presidential powers as part of Russia probe". Of course, many organisations did point out there was no factual basis for the claims. However, because of its perceived news value, an unsubstantiated claim (one Trump no longer makes) was given credibility by partisan media (like Fox News) as well as a leading international public service broadcaster (see Cushion and Thomas 2018).
In 2016, the EU referendum and the Trump campaign represented a particular challenge to broadcast news standards, creating moments that put notions of impartiality and accuracy in conflict with one another. Research on the broadcast coverage of the EU campaign, for example, found broadcasters often reported the tit-for-tat exchanges between THINK TANKS, TELEVISION NEWS AND IMPARTIALITY 481 politicians, with statistics traded back and forth, with little evidentiary commentary or mediation. This made it hard for audiences to assess where the weight of evidence lay (Cushion and Lewis 2017). The study also found that while broadcasters were successful in balancing Leave and Remain points of view, they did not feel compelled to underscore this with a sense of party political balance. Because divisions within the ruling Conservative Party were regarded as newsworthy, Conservative voices dominated coverage on broadcast news, outnumbering centrist or left-wing perspectives by a factor of four to one (Cushion and Lewis 2017).
These points raise important and critical questions about the practice and maintenance of "due impartiality" and accuracy in broadcast news. There is, however, a lack of clarity about what constitutes accuracy or impartiality, when these ideas should be put into practice or how they should be measured. The Chair of the BBC Board, David Clementi, has proposed the use of a scientific approach to the measuring of impartiality (cited in Shermin 2017). Such an approach would appear to be both inevitable and necessarynot only to offer critical scrutiny but to defend broadcasters from partisan attacks when the evidence does not support it. There is, however, some resistance amongst practitioners to this idea. So, for example, when the research group Media Tenor reported findings suggesting that over a 15-year period, BBC coverage of the European Union had been significantly more negative than positive, the BBC responded with the following statement: "It's just not possible to measure impartiality through some sort of mathematical formula. BBC News reports on the European Union fairly and impartially and we're satisfied our coverage achieves a proper balance" (cited in Jackson 2016). It is important to note that rather than offer a methodological challenge to Media Tenor's research (what constituted "negative" or "positive" coverage, for example), this rebuttal dismissed the very notion of applying scientific forms of measurement to impartiality. The rebuttal is thereby undermined by its own logic, rejecting a claim based on content analysis, yet asserting impartiality based on a vague reference to evidence.
Since the BBC has given a great deal of thought to the meaning and interpretation of due impartiality, this rejection needs to be taken seriously. Nonetheless, we would argue, to reject the use of any independent criteria with which to make judgements is ultimately untenable. Without some form of measurement, the absence or presence of impartiality becomes entirely subjective-simply a question of news judgement about a particular event or topic without any systematic means of verification. The basis upon which the BBC was "satisfied our coverage achieves a proper balance" is entirely absent. Indeed, broadcasters often do rely on broad mathematical forms of measurement to achieve fairness and balance over a period of time or an event (such as election campaigns-see Cushion and Thomas 2018), suggesting that impartiality is, in part, something that can be measured and quantified.
In 2007, the BBC Trust published a review that sought to redefine impartiality in a world where opinion did not always fall into two camps. It concluded that Whereas opinion used to be balanced in simple alternatives-and could be measured in tilts of the seesaw or swings of the pendulum-nowadays a more appropriate metaphor might be the many spokes of the wagon wheel … One opinion is not necessarily the exact opposite of another, nor do they all reach the extremity of available argument (Bridcut 2007) Five years later, the Trust commissioned Cardiff University to undertake a breadth of opinion study to explore whether the impartiality of BBC coverage had, indeed, moved from a "see-saw" to a "wagon wheel" approach (Wahl Jorgensen et al. 2016). Comparing BBC news coverage in 2007 and 2012, their study found that a fairly traditional notion of impartiality-the see-saw approach, or the "paradigm of impartiality-as-balance"remained firmly in place (Wahl-Jorgensen et al. 2016, 15). This is, perhaps, not surprising -the wagon wheel metaphor is far more difficult to operationalise than more binary models of impartiality.
The Cardiff University review also raised questions about the way in which news values shaped the nature of the binary views on offer. So, for example, debates about the European Union tended to reflect a parochial political debate in which the European Union was generally seen as a problem. Since domestic politicians have more incentive to blame the European Union than give it credit, there were few voices articulating the social, economic and environmental benefits of EU membership or the European Union's positive achievements (Wahl Jorgensen et al. 2016).
One of the key questions that underlies discussions of fairness, accuracy and impartiality is an assessment of when to operationalise these ideas. If the weight of evidence falls on one side rather than another, a commitment to accuracy runs counter to crude notions of balancing. So, for example, broadcasters have been accused of giving equal weight to claims about anthropogenic climate change (Boykoff and Boykoff 2004) or to unfounded claims about the MMR vaccine (Lewis and Speers 2003), thereby distorting a high degree of scientific consensus and failing to reflect the weight of evidence. Indeed, in response to a BBC Trust review on their coverage of science, the BBC shifted its editorial position on the coverage of climate change towards one that acknowledged high levels of scientific agreement.
Most of the academic research on impartiality and bias has tended to focus on areas where there are fairly simple binaries, defined less by the weight of evidence than by a conflicting set of ideas or interests. So, for example, the Glasgow Media Group's work has focused on the binaries between the views of workers/trade unions and employers/ business leaders (Glasgow Media Group 1995) or between Israelis and Palestinians (Philo and Berry 2004).
The most long-standing conceptual binary underpinning notions of impartiality in the coverage of public affairs is a political continuum between left and right. Like all binaries, it has its limitations-some ideas do not fit neatly on a left-right spectrum-but it nonetheless encapsulates a wide range of political divisions and is often articulated through party political proxies, most of whom can be positioned somewhere along a left-right continuum. While much of the research explores areas which relate to a leftright axis, there has sometimes been a reluctance to use it as a primary gauge to measure impartiality. So, for example, the BBC Trust's impartiality reviews have covered a range of topics including business, religion, immigration, the European Union, rural affairs, devolution and science. A review explicitly focused on broader questions of political or ideological impartiality is a notable absence from this list. This is, in part, because of the sensitivity surrounding these questions.
So, for example, the Trust's review of business coverage avoided some of the broader ideological questions about the relationship between business and other interests (BBC Trust 2007). The report's desire for a more consensual approach meant that it tended to assume that business coverage was politically neutral, glossing over the different interests THINK TANKS, TELEVISION NEWS AND IMPARTIALITY 483 of producers and consumers (consumers want the best product at the cheapest possible price, while businesses want to spend as little on the product as they can and sell it for as much they can), as well as the differences between business owners and employees. So while the report acknowledged that "audiences are served in their identity as consumers. But they are not that well served in their role as workers" (BBC Trust 2007, 9), it did not address the important questions of political impartiality raised by this point. Similarly, Cardiff University's research for a Trust review of the coverage of a range of issues found, by a series of measures, that "Conservative dominance in 2012" of BBC news was "by a notably larger margin than Labour dominance in 2007" (Wahl-Jorgensen et al. 2013, 5). While this would seem to raise very basic issues of political impartiality, it was not explored in the Trust's subsequent report.
In this article, we build on these research studies and tackle the issue of political impartiality head on. There are, of course, many binaries embraced by the principle of impartiality, but because of its abiding centrality to contemporary politics, we would argue that the best place to begin a scientific analysis is to explore the extent to which news coverage is broadly balanced on a left-right political spectrum.
Hopmann, Van Aelst, and Legnante's (2012) review of the field identified three main areas where this kind of political balance has been established or scrutinised in media coverage. These have all tended to focus on those aspects of news coverage that best lend themselves to coherent systems of measurement. First, media visibility, assessing when (and for how long) different actors are represented. Second, tone, when coverage is evaluated according to whether it is favourable or not towards an actor. Third, issue balance, where the agenda of news reporting is seen to favour one actor/political standpoint over another.
When possible, a systematic approach to impartiality should explore all three of these dimensions. In this article, however, we want to focus on media visibility, in order to scope out an approach based on a left-right political spectrum. In this study we wanted to focus on a particular set of actors-think tanks-which play a role in shaping our understanding of public affairs, which may represent a political perspective but which are not so obviously defined in political terms.

Thinking About Thinks Tanks
Our study sought to establish the broad ideological range of think tanks that informed BBC programming in 2009 and 2015. Think tanks represent an important source of knowledge and expertise in most Western democracies (see Kelstrup 2016). They can influence the formation of policy, set media agendas and inform public opinion. According to Ahmad (2008, 534), think tanks constitute "an institute, organisation, corporation or group that conducts research and engages in advocacy in areas such as social policy, political strategy science or technology issues, industrial or business policies, or military advice". As they have grown in size and scope over recent decades, scholars have sought to develop typologies to classify think tanks according to their aims, structures and affiliations. So, for example, McGann and Weaver (2000) developed a typology of four types of think tank-academic, contract, advocacy and party (political)-according to their funding, agenda, ideological orientation and research goals.
Our interest in think tanks is part of a broader analysis of the ideological influence of news sources and the extent to which they inform media agendas over time, in different political contexts and news programming. While many think tanks are not overtly ideological, the fact that a significant number of them do lean to the left or right provides us with a fairly straightforward yardstick: just as impartiality requires broadcasters to give parties on the left and right broadly equal treatment, we would expect an impartial broadcaster to give roughly equal weight to right-leaning and left-leaning think tanks.
Much of the research about think tanks and ideology comes from the United States, since partisan organisations have become a more prominent part of its political culture. These have been termed "advocacy think tanks", since they actively seek to promote a particular ideological agenda. Medvetz's (2012) comprehensive analysis of US think tanks identified a rise in well-resourced conservative think tanks since the 1970s, with an ability to lobby political elites and shape policy agendas. Their influence has also been traced in media coverage. Rich and Weaver (2000, 99), for instance, examined six US newspapers between 1991 and 1998 and discovered "ideological biases … benefiting conservative think tanks in the Washington Times, Wall Street Journal, and think tanks of no identifiable ideology in the New York Times, even after controlling for budget size". Likewise, McDonald (2014) examined television, press and radio news reporting of think tanks specialising in education in 2001 and 2006. She concluded that "conservative think tanks produced the largest number of education media citations, followed by centrist think tanks. Liberal/progressive think tanks and university-based education-policy centers had little to no media presence" (McDonald 2014, 845).
The FAIR (Fairness and Accuracy in Reporting) analysis since 1996 represents the most comprehensive picture of the presence of think tanks in US media coverage over time (Dolny 2013). Its annual reviews (up until 2012) have shown that while centrist think tanks make up the majority of citations, right-leaning think tanks consistently gain a greater share in media coverage than their left-wing equivalents. According to its 2013 report (which examined major newspapers, radio and television news; Dolny 2013), 46 per cent of think-tank citations were centrist, 35 per cent were right-leaning and 19 per cent left-leaning.
While the dominance of conservative-based think tanks in US media is well established in the literature, the political map of think tanks in other countries (with different media systems) is less clear (see Kelstrup 2016). Indeed, when compared to many other media systems, the United States is often viewed as exceptional (Curran 2011;Pickard 2014), since it is primarily market-driven, lightly regulated and, in recent decades, increasingly partisan (Stroud 2011). While the United States has shifted away from broadcast regulations on fairness and balance (Strömbäck and Lee Kaid 2008), many European countries have maintained regulations about the impartiality of broadcast news coverage of politics and public affairs (Cushion 2012). However, there are ongoing debates in many advanced democracies about how to police the impartiality of broadcasting in an increasingly unregulated online and social media environment. Interestingly, research in the United States also suggests that the success of right-leaning think tanks is not simply a matter of the growth of right-wing broadcasters (like Fox): the resources available to right-leaning think tanks like the Heritage Foundation, the American Enterprise Institute and the Cato Institute also give them a prominence in traditionally non-partisan media.
The literature about political balance (Hopmann, Van Aelst, and Legnante 2012) also reveals an "incumbency bonus" where ruling parties attract greater coverage-both of politicians and allied organisations (such as the more partisan think tanks). Research by Wahl-Jorgensen et al. (2013, 15-16) in the United Kingdom found a version of the "incumbency THINK TANKS, TELEVISION NEWS AND IMPARTIALITY 485 bonus", but one skewed to the right. Their study found that while the Labour and Conservative Parties were represented in roughly equal proportions when Labour was in power in 2007, when a Conservative-led coalition was in government in 2012, the right-wing party received significantly more media attention. In other words, the ideological balance of party political coverage favoured the Conservative Party when they were in power and granted them equitable status with the ruling party when they were in opposition. Dolny's review of the coverage of think tanks in the United States, however, found that -with think tanks at least-this can work the other way, suggesting that "media may be more open to a progressive economic critique when a Democrat is challenging Republican policies than when the Democrat is defending his own record" (Dolny 2013).
In order to test the impact of incumbency, we chose two sample years when a centreleft party (Labour in 2009) and centre-right (Conservatives in 2015) were in power. In 2009, New Labour had been in power for 12 years, and, as Schlesinger (2009) has documented, Labour's policy agenda was heavily influenced by the rise in more liberal and left-leaning think tanks. By 2015, however, the Conservatives had been in power (albeit leading a coalition government) for five years (later that year they would become the sole governing party after a General Election). According to Hartwig (2013), a number of right-wing think tanks helped to develop the Conservative Party's policy agenda when in opposition and power.
Our study explores whether coverage of think tanks is more ideologically balanced in the United Kingdom's more regulated, impartial broadcasting system than it is in the United States. We chose to focus on BBC news programming because of its prominence in the UK news ecology and its reputation for editorial independence and impartiality. Political parties, of course, have relatively clear ideological credentials, whereas the agenda of think tanks can be more elusive and difficult to determine. Scholars-primarily in the United States-have developed typologies to classify the type and ideological aims of think tanks (McGann and Weaver 2000). But in the United Kingdom only one study to date has sought to label think tanks on a left-or right-wing political spectrum. The Centre for Policy Studies used a proxy measure to determine a think tank's ideological leanings, based on how often they appeared in right and left newspapers over the same period of analysis. In doing so, they categorised left-and right-wing think tanks according to whether they were cited in the (left-leaning) Guardian and the (right-leaning) Daily Telegraph.
While this produced a list that correlates very roughly with political affiliation, it features a series of anomalies, some of which are significant. While newspapers clearly have political preferences, they are indirect, second-hand filters for assessing the political leanings of sources, especially if no account is taken of the context in which they appear. So, for example, the Centre for Policy Studies' list categorises Chatham House as one the most left-wing think tanks in the United Kingdom, puts the Global Warming Policy Foundation (an outlet for climate change sceptics) on the left side of the ledger and the Institute for Public Policy Research (IPPR; a think tank associated with the centre-left) on the right. Our more rigorous methodological approach-as well as conventional wisdom-indicates that these classifications are untenable: Chatham House being unaffiliated, the Global Warming Policy Foundation right-leaning and IPPR left-leaning. While the Centre for Policy Studies method has the merit of simplicity, a more robust approach cannot avoid looking at the think tanks themselves and applying clear criteria to classify them in political/ideological terms.

Sample and Methodology
We used the BBC's Redux archive to identify when and where think tanks appeared in BBC coverage. This consisted of over 30,000 BBC television news and current affairs programmes-the bulk of which were broadcast on the BBC News channel or BBC One in the years 2009 and 2015 (see Table 1).
We then consulted a wide range of sources to identify think tanks based or operating in the United Kingdom. This allowed us to compile a list of 134 research institutes/think tanks (see Appendix A). After entering the names of all 134 think tanks into Redux, we found that exactly half (67) appeared in or were quoted on BBC news programmes in 2009 and 2015. We then sought to classify this smaller group by political leaning.
Our approach to classifying the political leaming of think tanks was adapted from Andrew Rich's (2004) comprehensive analysis of US thinks tanks. This involved five elements: 1.
In the first instance, we looked at a think tank's self-description, since some think tanks place themselves on a political spectrum (e.g. the Fabian Society and Compass on the left and Bright Blue and the Bow Group on the right). Most think tanks, however, do not overtly self-identify.

2.
We then looked for key words or phrases associated with left/right positions, so, for example, an emphasis on "fairness", "equality" or "progressive" ideas tends to be used by left-of-centre think tanks, while an emphasis on "free markets" is associated with right-of-centre think tanks in the United States. This helped us to categorise think tanks like the Institute of Economic Affairs, the Adam Smith Institute, the Centre for Policy Studies and Civitas on the right, and the IPPR, the New Economics Foundation, the Policy Network and the Intergenerational Foundation on the left.

3.
We combined this with an identification of ideas that tend to be associated with left or right positions, e.g. the desire to reduce immigration (right), support for asylum seekers or refugees (left). So, for example, the Tax Justice Network is an international organisation campaigning against corporate tax avoidance, a cause which tends to be associated with the left, while The Scientific Alliance and the Global Warming Policy Foundation campaign against or question the scientific consensus on climate change, a position now associated with the political right (Whitmarsh 2011).

4.
We examined the political backgrounds of members of advisory boards or trustees.
Where there was a preponderance of members affiliated with political parties on the right or left, we categorised the think tank accordingly. So, for example, The Bruges Group (coded as right-leaning) is run by a Conservative Councillor Robert Oulds, and has two prominent Conservative politicians as President (Lord Tebbit) and Vice President (Lord Lamont). The Electoral Reform Society (coded as left-leaning) has a Deputy Chief Executive who is a former Labour politician in New Zealand and its Council and Advisory Committee contains many Labour and Liberal Democrat politicians and one Green, but no Conservatives.

5.
We also looked at non-politicians with clear political associations involved in the organisations. So, for example, the Director of Policy Exchange is formerly the Chief Leader Writer on the Daily Telegraph, a newspaper with a long tradition of supporting the Conservative Party (categorised as right-leaning), while the Equality Trust's Board Members include Zoe Williams of the Guardian and a former Senior Policy Officer at the Trades Union Congress (categorised as left-leaning).
There are undoubtedly grey areas between our non-partisan and left/right-leaning categories. So, for example, the Royal United Services Institute (RUSI), like many think tanks, bills itself as objective and non-partisan, and is not tied overtly to any political party. There is some evidence of non-partisanship: its chairman, former Conservative leader William Hague, took over from Labour peer Lord Hutton, and it has long-standing establishment credentials. It is, however, firmly rooted in the defence and security establishment, and there is a preponderance of Conservative voices in its governance and structure. There are two Conservative lords and one Liberal Democrat on its Advisory Council, but no Labour members. The testimony on its home page contains endorsements from four senior Conservatives, a US Republican (Henry Kissinger), former Directors of the UK Intelligence Service and the Central Intelligence Agency, the Queen and the King of Jordan. Many of its positions (e.g. in support of strategic alliances with Gulf States) put it at odds with human rights groups like Amnesty and certainly with groups like the Campaign Against the Arms Trade. For this reason, we categorised RUSI as right-leaning.
On the other side of the ledger, the Institute for Government appears to be similarly non-partisan, but while its board contains figures from across the political spectrum, its centre of gravity is moderately centre-left: its Chairman is Labour peer Lord Sainsbury and the board includes two other Labour politicians-Baroness Amos and Liam Byrne, a Liberal Democrat and a crossbench peer (Lord Sharkey and Lord Currie) and just one Conservative, Lord Finkelstein.
We appreciate that many on the right would see the RUSI as non-partisan, while many on the left would say the same about the Institute for Government or the Electoral Reform Society. Equally, some might question the non-partisanship of those we categorised as non-affiliated/centrist. So, for example, the Institute for Fiscal Studies (IFS) has been criticised from the left for some of its positions on taxation, 2 while during the EU referendum it received flak from the right for its assessment (in line with most economists) of the economic costs and benefits of EU membership. Others have argued that because the members of the IFS are not macro-economists, they tend to focus on balancing budgets. This can be problematic when parties of the left pursue a New Keynesian macro-economic strategy which favours using public investment to stimulate growth (and hence increase the size of the budget in the future). 3 The fact that criticisms come from left and right is not a justification for classifying the IFS as non-partisan (this assumes both criticisms are of equal weight), but our criteria are not finely tuned enough to classify it as right-or left-leaning.
Our method is, in this sense, limited to clear, verifiable criteria. While this may have its disadvantages (if think tanks keep their partisanship well-hidden, or where ideological bias is only established through a forensic examination of its output), it has the advantage of being replicable and transparent. Our results are outlined in Table 2.
At first glance, Table 2 suggests a remarkably high degree of even-handedness in BBC coverage, with news programmes showing a preference for centrist or non-partisan think tanks, with equal numbers in the left/right columns. However, when we look at the number of times each think tank is mentioned over the course of each year, a more precise picture emerges. We counted the number of times a think tank was mentioned during programmes. Table 3 shows that BBC programmes tend to favour non-partisan or centrist think tanks, a finding partly explained by the widespread use of the IFS-by far the most used think tank across the BBC (see Tables 4 and 5, which list the top 10 most cited think tanks in each category). Indeed, such was its predominance, that the IFS alone constitutes 47 per  cent of all references to think tanks in 2009 and more than half the total number of references (54 per cent of 2100) in 2015. Table 3 also indicates a clear shift to the right in the BBC's choice of think tanks. In 2009, there was a broad balance between left and right think tanks-left-leaning think tanks receiving 16.5 per cent of think-tank references and right-leaning think tanks receiving 18 per cent. In 2015, references to right-leaning think tanks remain at a similar level (16.2 per cent) while references to left-leaning think tanks are halved to just 8.5 per cent. To put it into a broad political context, when Labour was in power, the BBC's use of think tanks was relatively even-handed, but when a Conservative-led coalition was in power, the centre of gravity shifted to the political right.
This finding echoes one of the few other studies to make direct comparisons between recent Labour and Conservative periods in office. Wahl-Jorgensen et al. (2013), in their impartiality review for the BBC Trust, compared BBC coverage in 2007 (Labour in power) with 2012 (a Conservative-led coalition in power). They found fairly equal  (Tables 4  and 5), we see some continuity-the popularity of the IFS, Chatham House and the RUSI is consistent across both years. In keeping with the rightward shift, however, we see a significant fall in the use of some left-leaning think tanks. In 2009, the three most cited left-leaning think tanks-the IPPR, the New Economics Foundation and Compass-comprised, between them, almost 10 per cent of think-tank references. By 2015, all three remained active but had almost disappeared from BBC news, with around 1 per cent of think-tank references between them.
We should note that these findings are at odds with the Centre for Policy Studies research, which suggested that because the BBC's choice of think tanks correlated more closely with the Guardian's than the Telegraph's, it demonstrated that the BBC's centre of gravity is on the left. This disparity is largely explained by their methodology, which we described earlier. So, for example, their classification of the oft-quoted Chatham House as very left-wing significantly skews their data-so while Chatham House may have received more coverage in the Guardian, by any independent criteria they are either centrist or non-partisan.
Perhaps surprisingly, however, when we tested the Centre for Policy Studies' study rankings in our 2015 data-set, we found that it suggested a more dramatic right-wing bias than our own. If we split its list into roughly equal thirds (14 right-leaning, 13 non-partisan, 14 left-leaning) and map it across our data, we find right-leaning think tanks constitute 69 per cent of all think-tank mentions, compared to 13 per cent on the left. Their ranking of the IFS as a right-leaning think tank-given its popularity on the BBC-clearly distorts the picture, but even if the IFS is removed, the Centre for Policy Studies' list still produces 38 per cent of think-tank references on the right and 25.5 per cent on the left. Although this may suggest that its findings reflect our study, for reasons already outlined we consider that our own method of political classification is more robust. Table 6 shows that the 2015 imbalance (between left and right references to think tanks) differs across the many BBC outlets. BBC Four (which has one regular news programme, World News Today) is the most balanced, although the numbers here are small. BBC One, the BBC's most popular channel (which features its flagship news bulletins) favours centrist or non-partisan think tanks and has a less significant rightward tilt than the BBC News channel, which combines news and current affairs programming. News/current affairs programmes on BBC Two very clearly favour right-wing think tanks. This is mainly accounted for by two current affairs programmes, Daily Politics and Newsnight, where the ratio of right-leaning to left-leaning think tanks is a striking six to one. In other words, while most BBC outlets generally favoured right-wing think tanks, this was least pronounced in conventional broadcast news bulletins and most dramatic in current affairs programming.

Sourcing Think Tanks: Rising to the Challenge of a Partisan News Environment
Our findings indicate that when a Labour government was in power, the BBC was generally even-handed in its use of think tanks-favouring centrist or non-partisan organisations and giving fairly equal weight to those on the right and left. In the more recent period of Conservative rule, centrist or non-partisan think tanks are still preferred, but the use of left-leaning think tanks drops significantly.
Our findings suggest that, overall, the United Kingdom has not mirrored the United States in reflecting a think-tank culture where many of the big beasts (such as the American Enterprise Institute, the Cato Institute and the Heritage Foundation) are on the political right. Most of the most quoted think tanks on the BBC are centrist or non-partisan, while the list of think tanks used by the BBC across both periods is remarkably balanced. In this sense, the BBC appears acutely aware of its commitment towards impartiality, largely drawing on sources that are not ideologically disposed to left-or right-wing perspectives. Nonetheless, in 2015, we found a subtle but significant shift towards favouring right rather than left-leaning think tanks, most strikingly in current affairs programming.
Overall, we would argue that our findings should be understood in the context of a number of recent studies, all of which suggest that the BBC has, in recent years, moved its centre of gravity to the right. As we have indicated, the comparative study of Wahl-Jorgensen et al. (2013,2016) showed relative even-handedness in terms of party political representation (indeed, a small degree of pro-Labour bias) when Labour was in power, switching to a clear Conservative Party dominance when the Conservatives were in government after 2010. Other studies of more specific issues or time periods reinforce this picture. Berry's (2016) research on the BBC's economic coverage found that the BBC replicated a Conservative pro-austerity narrative following the 2008 economic crisis, while a study examining television news coverage of the 2015 General Election campaign identified that the BBC was more likely to follow the Conservative rather the Labour Party's issue agenda and was influenced by the editorial agenda of right-wing newspapers (Cushion et al. 2016a). Most recently, research about television news coverage of the 2016 EU referendum (Cushion and Lewis 2017) found that while the BBC-and broadcasters in general-were scrupulously even-handed in representing the Leave and Remain campaigns, they significantly favoured right-wing sources during the campaign.
While none of these studies are conclusive, they do suggest a pattern. Since the BBC is clearly committed to impartiality, how can a right-wing shift in the sources used by news and current affairs programming be explained? A closer look at the financial resources or public relations strategies of think tanks-beyond the scope of this study-may provide some insight. But one visible influence on BBC programming over recent years has come from the bullish and campaigning right-wing UK press (notably the Mail, the Express, the Telegraph and the Sun) who dislike public service broadcasting for both commercial and ideological reasons (Lewis 2014(Lewis , 2015. There is no equivalent pressure coming from the left (the United Kingdom's two left-leaning newspapers, the Guardian and the Mirror, do not put pressure on the BBC). Following the 2017 General Election when-in spite of fiercely negative press coverage-the Labour Party won a far greater share of votes than many predicted, several commentators suggested the rise of new news sites and sharing platforms such as Facebook have diminished the influence of the right-wing press (Jackson 2017). While the United Kingdom's media ecology is clearly changing, we would argue it is premature to suggest the agenda-setting power of right-wing newspapers is over. Newspapers, after all, continue to inform broadcast programming and are widely read by journalists, policy makers and politicians. Indeed, while the broader media ecology has become more diverse, broadcasters continue to privilege the UK press (in ways that reflect rather that filter their biases) as a source of content and opinion (see Renton and Schlosberg 2017).
In our view, public service broadcasters must rise to this challenge by maintaining a commitment to impartiality by keeping a distance from-rather than too easily reflectingpowerful partisan influences; or, to put it another way, to be impartial requires a clear understanding of the lack of impartiality elsewhere, and a more independent awareness of how stories emerge, and how issues and sources should be balanced and explained. More closely monitoring which think tanks inform their programming is one way the BBC-and other broadcasters-can more carefully safeguard impartiality in an increasingly partisan media environment.