Content Analysis Preliminary Findings: Part II, Partisanship
This is the second of three blog posts which outline some preliminary findings from our content analysis of alternative media in the UK.
In this post, we provide an overview of how the media outlets in our study cover political parties and how they talk about them, which we use to infer some conclusions about partisanship. Part one, on agendas, can be read here. This post contains the sample we worked with, but as reminder, we analysed articles from four periods of time (6 – 25 October 2015; 9 – 29 October 2016; 30 April – 7 June 2017 [General Election]; and 8 – 28 October 2018). Once again, prior to peer-review, this is not intended as an in-depth or final account.
We began by exploring the extent to which these outlets focused on particular political parties. As our previous post showed, party politics in general makes up the predominant focus of most of the outlet’s coverage. That is, articles that are about party politics. The graph below shows how many articles saliently focused on a particular political party across the full range of coverage. (By salient, we mean that within an article, a party or individuals within it were a key part of the narrative. We only coded one party per article, meaning we made judgements based on primary salience where more than one party was discussed).
As this graph shows, both The Conservative Party and the Labour Party receive the most attention by far. Across the full sample, 64.8% of all articles focused on a particular party and of these 34.6% and 54.4% focused on the Labour Party and The Conservative Party, respectively. None of the outlets give much attention to smaller parties, with the small exception of UKIP.
The reasoning here, perhaps, is a combination of both a ‘hard’ political agenda and an adherence to political conflict as a news value, which of course parliamentary, two-party politics induces. But also notable, and perhaps overlooked, is the amount of content, over a third across the sample, which does not substantively focus on any particular party.
These results may be partly skewed by the inclusion of our election sample, which makes up 50% of the content we analysed. So here’s how this focus breaks down if we remove the election subsample and just look at content outside this period.
As you can see, the pattern is broadly the same, although there are some shifts that warrant closer analysis. A biggest shift is in The Skwawkbox’s coverage, which focuses more more on The Labour Party and much less on The Conservative Party outside of the general election. Also noticeable is the consistent balance of coverage that Guido Fawkes provides to both the Labour Party and the Conservatives.
This evidence tells us about how often these outlets cover particular parties. But what about sentiment? To understand this, we analysed how whether the coverage given to a particular party was supportive, critical, neutral. Because the sample breakdown begins to get a bit more complicated at this point, the following table specifies the number and percentage of articles produced by each outlet across the full sample that we are discussing. For example, 57% of all articles produced by The Canary within our sample are critical of a political party.
As this table shows, when most outlets cover political parties they chose to present them negatively a lot of the time. The exceptions are Breitbart London, with 25% of its overall coverage featuring neutral presentations of political parties. (This is an interesting finding; Breitbart London’s content often features a heavily objective, neutral style but appear to overall report stories that suit its editorial agenda).
To colour this in, we broke down these segments of data by party. The following graph shows the proportion of critical coverage given to each party.
Clearly, we can see how the left-wing outlet’s content is aligned to their editorial values of anti-conservatism. On the right and with Guido, we see some more variation - they are perfectly willing to be critical of the Conservatives. Our experience with the data suggests much of this is Brexit-related attacks on the leadership. For Guido, it probably reflects their generally antipathic approach towards politicians.
Following on, this graph displays the percentage of content that is supportive of a political party.
Again, pretty sharp results.
These two graphs appear to confirm the “hyperpartisan news” claims put forward. We’d urge caution here without further analysis, for two good reasons. Firstly, the there is a strong degree of content in the full sample that is not party related. This complicates the picture. Secondly, the label “hyper-” doesn’t mean much without comparison with the wider media system. Recent research by Loughborough University’s Centre for Research in Communication and Culture (using a different methodology and sample period), for example, demonstrates the high degrees of partisanship in the British press. So it is important to be cautious.
Moreover, it is so (so) important to contextualise these findings with the volume of output and our dataset. We are talking fairly small and uneven subsamples once broken down. For example, in these two graphs, our analysis of Novara Media based on just 11 articles (so not especially reliable) compared to over 500 for The Canary (more reliable). These are of course dealt with in peer-review.
Overall, however, this begins to provide a clearer empirical picture when it comes to alternative online political media and their partisanship. As our findings and analysis develop, and as more research is published (particularly studies comparing mainstream and alternative media), we hope to be able to sharper this picture.
