Comparing the election coverage in The Australian and The Age, you’d think the campaign was happening in two parallel universes.
You’d also be confused about its value as news.
A content analysis of the front pages of these two newspapers up to the end of the campaign’s fourth week (Friday, 9 November) shows The Australian ran 72 front-page items on the election, and The Age a mere 32.
Aside from sheer quantity, however, there is a dramatic difference in the partisan biases of the coverage. The table shows the percentage of each newspaper’s items that fall into various categories of bias, from “clearly positive for the Coalition” to “clearly positive for Labor”.
Clearly positive for the Coalition
Marginally positive for the Coalition
Marginally negative for the Coalition
Clearly negative for the Coalition
Clearly positive for Labor
Marginally positive for Labor
Marginally negative for Labor
Clearly negative for Labor
*Percentages add to more than 100 because some stories count as both positive for one side and negative for another. For example, an attack by Howard on Rudd, without a response from Rudd, counts as positive for the Coalition and negative for Labor.
Four findings stand out:
- The Age was much more likely to run neutral items (15.6% to 9.7%).
- Although both papers showed bias, The Australian showed more.
- The direction of The Australian’s bias was clearly in favour of the Coalition and against Labor.
- The direction of The Age’s bias was clearly in favour of Labor and against the Coalition.
This is obvious from a simple calculation of the net bias ratings, based on the total positive minus the total negative score for each party.
The Australian’s net bias rating for the Coalition was positive 19.6, and for Labor negative 9.7.
The Age’s net bias rating for the Coalition was negative 15.6 and for Labor positive 3.1.
Of course, all content analysis is subject to bias by the analyst, no matter how hard he or she tries to be neutral, and so it is only proper that I should declare that I intend to vote Labor at this election. I should also declare that from 1986 to 1993 I was an Associate Editor of The Age.
This analysis is based on the stories selected for page one, the language and subject-matter of the headline and first paragraph, and the order in which the various components of the stories appeared. Main page-one pictures are counted; small pictures and cartoons are not. The items are not weighted for relative prominence.
The qualitative differences between the papers were as striking as the quantitative ones. A few examples:
Labor caught out again on money
Now the PM flunks the numbers test
Labor at war over IR laws
PM, Rudd warned on spending
Basic error in Labor tax plan
Rates threat as tax-cut billions flood economy
Howard to face nation of workers (indicating good jobs figures)
Three rate rises ahead, says ANZ
Garrett Kyoto blunder
Howard’s tech school revival
These few examples — among many — illustrate extraordinary differences in the assessments by the two newspapers of what the important stories were on various days through the campaign.
Putting it all down to party bias is just silly, because although they are both broadsheets, they appeal to slightly different markets – The Australian to a more business-oriented national audience and The Age to a more general Victoria-based audience.
Even so, the differences are striking, and suggest news judgments are influenced to some extent by the papers’ worldview and probable political preferences. We will probably be able to judge this when they publish their election leaders next week.
An entertaining sidelight in all this is what might be called the Shana factor. This refers to Dennis Shanahan’s efforts in The Australian to put the most positive spin on Newspoll results for John Howard.
Take his page one report on 3-4 November. Here he presented the findings from Newspoll’s survey of the 18 most marginal Coalition seats in the four states of New South Wales, Victoria, Queensland and South Australia.
He leads by saying Howard has fought back in NSW and Victoria, which is all very well except the figures suggest a loss of between 1 and 11 seats for the Coalition in those states and no gains.
It is not until the eleventh par that he gets around to telling us that the two-party-preferred vote across the 18 seats is 53-47 to Labor. This difference of six percentage points would, if repeated on election day, deliver 12 of these 18 seats to Labor.
In Queensland, the swing is averaging 9.6% against the Coalition and this, if repeated on election day, would deliver 10 more seats to Labor. We don’t learn about the Queensland swing until paragraph nine.
Compare this with the analysis by Sol Lebovic, the founder and former chairman of Newspoll, on page 6 of the same day’s paper. Under the heading “Little for Coalition to cling to”, he tells us within the first six pars that the average 7% swing across the 18 seats would, if repeated on election day across those four States, deliver 25 seats to Labor, giving it an easy win.
This, by any conventional set of news values, is the lead of the story.