Media Slant in Walkley

The Walkley Magazine has published an article by myself and Andrew Leigh on media slant. It is reproduced over the fold.

True blue view from the red centre

Joshua Gans and Andrew Leigh

The Walkley Magazine, Issue 58, 14th October 2009, p.50

It is a hardy perennial of public debate. Politicians or their shills attack a media outlet because they believe it to be slanted against their cause. In the US, this has been notable in Republican attacks on the Washington Post and New York Times for a left-wing slant. Indeed, the whole ‘fair and balanced’ positioning of Fox News was to counter that perception.

In Australia, these issues also arise occasionally. The most recent was an accusation just last month from former Treasurer Peter Costello against the ABC for what he saw as a long-standing bias against his political views. “With the ABC the line of questioning is always predictable. It always comes from the Labor/Green perspective,” he wrote. By comparison, he argued, the ALP received a free ride. Costello concluded: “I am not now at the mercy of the media so I can afford to say what everyone on the conservative side of politics knows: the ABC is hostile territory.”

But are such perceptions true? The alternative hypothesis is that the ABC is doing its job and giving a hard time to any politician who fronts it for an interview. Let’s face it, under a well-informed grilling, it’s easy to imagine thinking that the interviewer might secretly be a mole for the other party.

A number of US studies have set to explore media slant in that country. Instead of looking at isolated instances, these studies have sought a more ‘objective’ test that covered a large volume of news outlet reporting. In one carefully designed study that built upon the US’s flourishing think tank population, researchers Tim Groseclose and Jeffrey Milyo sought to identify a coincidence of citation. First, they looked at which politicians referred favourably to particular think tanks. They then scored that in terms of citations relatively to one party or another. For instance, the Brookings Institution was cited more often by Democrats than Republicans and was given a score that reflected that.

Groseclose and Milyo then took this information and tested whether particular newspapers cited particular think tanks more often than others. The logic of this was that if it turn out that a newspaper cited think tanks who were also more heavily cited by politicians of a particular party, they could score the degree of slant of that outlet – relative to all other outlets or what should have occurred had think tank attribution been simply random.

In many respects that study led to the expected results. First, there was a significant degree of polarization amongst US news outlets. Some were right leaning while others were left-leaning. And they were the ‘usual suspects.’ But overall, media outlets in the US were to the left of the typical US voter. Subsequent studies by economists suggested that this was not the result of some strategy on the part of media barons but instead was ‘market positioning.’ You may not like Fox News but, in many respects, they were just filling a market gap given that much of the US media was to the left of the electorate. (If you have recently returned from a trip to the US and find this hard to believe, it is helpful to remember that the US electorate is also well to the right of the Australian electorate).

Given that similar perceptions arose in Australia, we sought to replicate the US research here. Australia does not have the same think tank population but we do have a set of public intellectuals that perform the same role and are regularly cited by politicians and the news media alike. So in principle, we can use public intellectuals as a “crosswalk” between parliament and the media. But what public intellectuals should we choose? After surveying our options, we eventually decided to start with a list created in 2005 by Michael Visontay for the Sydney Morning Herald. After supplementing Visontay’s list slightly, we had 155 individuals.

The list was further whittled down. To make the cut, you had to be favourably cited at least once in Hansard during the period 1996-2007. For example, when John Faulkner told the Senate, “Who do we finally have in the papers today? We get the real doddering fools like Paddy McGuinness, Piers Akerman and others trying to defend this government”, we do not count this as an ALP vote for Paddy McGuinness and Piers Akerman. (Incidentally, the public intellectuals who get criticised in parliament are mostly columnists.)

Of the 155 public intellectuals on our initial list, 48 received no favourable mentions in parliament, leaving 107. Among these, 21 public intellectuals were cited significantly more often by one side of politics than the other. Coalition politicians were significantly more likely to mention Marie Bashir, Geoffrey Blainey, Ron Brunton, John Hirst, Helen Hughes, Paul Kelly, Hugh Mackay, Wendy McCarthy, Noel Pearson, Ken Phillips, and Paul Sheehan. Labor politicians were significantly more likely to mention Larissa Behrendt, William Deane, Mick Dodson, Gerard Henderson, Michael Kirby, David Marr, Les Murray, Barbara Pocock, Anne Summers and George Williams.

Then it was time to look at media citations. Our 107 stars were cited over 84,000 times between 1999 and 2007. (By the way, you’ll note that that covers the period where Peter Costello was most prominent). This included newspapers, radio and television.

To benchmark media slant, we looked at each party’s overall disposition to cite public intellectuals favourably at all. 47 percent of all citations came from the Coalition so a score of 0.47 was our benchmark. If a media outlet cited those Coalition favourites more often than that, they were slanted towards the Coalition. Otherwise, they were slanted towards Labor. But our measure also allowed us to see how slanted they were. Yes, an outlet with a score of 0.48 is right-leaning but it is unlikely to be significantly so. The more times an outlet mentioned public intellectuals, the more precisely we were able to gauge its ideological position. In deciding whether to tag an outlet as slanted, we used the common social science rule that it needed to be statistically significant at the 95 percent confidence level.

And so now to the results. In contrast to their American counterparts, the Australian media is a pretty centrist punch. There are some more to the left, and there are some more to the right, but overall we couldn’t say most outlets were significantly slanted. This might be a disappointment if some outlets were trying to position themselves as slanted. But others, such as ABC’s Radio National, will probably take pleasure from our finding that they were smack bang in the middle of the ideological spectrum.

Of course, in this part of our study, there was one outlet that stood out as more slanted than randomness would suggest: and that was ABC Television News. And which way? It will surely surprise the former Treasurer to find out that all of these years ABC Television News was more for his side than against him. To be precise, ABC Television News is significantly more likely to quote the kinds of public intellectuals that Coalition politicians mention favourably in parliament than the kinds of public intellectuals that Labor parliamentarians cite.

Given the nature of this exercise, we did not stop there and conducted a few robustness checks using alternative methodologies. In particular, we conscripted a hapless quintet of coders to read all front-page articles from the 2004 election campaign and form an opinion as to whether they were slanted towards one party than another. We did the same for the article headlines. These results confirmed our main finding (that the Australian press is pretty centrist), although this time The Age stood out as left-leaning. (There is a certain irony in the fact that Costello used its opinion page to attack the ABC.)

If there is any place in the Australian media where you see substantial slant, it is not on the factory floor, but in the corner offices. During 1996-2007, 36 of the 44 newspaper editorial endorsements were for the Coalition. Over the period 1998-99 to 2006-07 (the years when Australian Electoral Commission data are readily available), media proprietors gave 39% more money (a difference of $158,145) to the Coalition than to Labor.

It has been interesting to watch the media and blogging commentary as our results were released. There were calls of “I knew it” with regard to the right-leaning evening news programs on our national broadcaster, as well as attacks on statistical methods being useless if they didn’t confirm “what everyone knew.” Our response is that the statistics are designed to test theories. If they don’t confirm your preconceptions, that might suggest that your bias really is one of perception rather than reality. Perhaps your experience is idiosyncratic, and not what you might have formed had you read every single report from Australian outlets over a decade. Only statistical super-crunching can get at that.

Why is the Australian media so centrist? One theory is that in the US, there is greater competition in the media market and more variety. That might mean that consumers are better satisfied, and perhaps also that the truth is being distorted. But these are matters for further study. For now, our hope is that star-crossed pollies think twice before they claim that the media really is out to get them.

Joshua Gans is an economics professor at Melbourne Business School and Andrew Leigh is an economics professor at the Australian National University. Both are contributors to the Core Economics blog at

Their study, “How Partisan is the Press? Multiple Measures of Media Slant”, is available at