Thursday, October 24, 2013

Project Proposal: Do Media Organizations with Shared Political Bias Similarly Cover Breaking News?


Student Information:
Michael Mori (MALD 2015)

I will be joining Trevor and Gustavo for the project “Connections amongst current Fletcher cohort,” and so am posting this proposal as an up-for-grabs SNA project for other students joining the second module of the course.

Research Question

Do major media organizations that have the same political biases also similarly cover breaking news events? (Sub-question: do political biases in media organizations affect their breaking news coverage, and if so to what degree?

Hypothesis

Media bias is reflected in breaking news coverage, but the bias is not as pronounced as it is in conventional coverage because of the need to cover relevant national attention grabbing stories; therefore, while we will see a correlation between media bias and breaking news coverage, it will not be a strong correlation and it will vary from outlet to outlet, despite evidence of shared bias.

Background

Anyone who regularly gets their news from a wide variety of online sources will have noticed biases in news coverage. These biases commonly take the form of political angles on particular stories and are reflected in the prioritization of coverage on the front page of the outlet’s app or website. Other times media organizations simply do not cover stories at all. However, when news is “breaking,” meaning that it is a big story takes precedent over all other stories under coverage at the given moment, most will give some coverage to it. Here’s where it gets interesting.

Many media organizations issue breaking news through alerts on mobile applications, via emails to subscribers and of course on their websites. Each of these organizations keeps a close eye on the coverage of their competitors and is responsive to their coverage, especially breaking news stories. I have access to a dataset of all the breaking news email alerts sent out by 15 major media outlets over the past 14 months.

This dataset includes a timestamp for each breaking news email, a start and end time for each breaking news event, subject tags for each email, top tags over various periods of time, number of alerts per week per outlet, number of breaking news events per week, and number of emails sent by each outlet on the each story as it evolved over time.

This data can be used along with information from numerous studies that have already been done on media bias to get a better idea of whether media organizations with similar biases have similar coverage of breaking news through email alerts.

Methodology

This project has four major steps of analysis. The first is to identify and select the best most recent study on the bias of each of the major media organizations for which I have data, and then to add this attribute to the data mix.

Once this has been accomplished, you should look through the breaking news events for the 14 months and categorize them in terms of bias; either as pro-liberal, pro-conservative or neutral (neither helpful nor harmful to either political group). Next, individually look at each media outlet and analyze whether its coverage of breaking news events reflects its alleged bias, and if so to what degree.

The fourth and most interesting step is to look at the organizations and see how they are connected to one another by media bias -- meaning whether or not they share the same bias -- and then look to see if and how their breaking news coverage overlaps. To do this, you will analyze the data to examine if media organizations with shared bias covered (or did not cover) the same breaking news events. Be sure to look into the extent of their coverage as the story progressed over time. Another point of inquiry would be into the use of tags. News sources that have a lot of tags in common and have a high overlap in coverage may be similar. The question is whether or not that similarity is reflected there alleged bias.

Potential Limitations


Social network analysis alone cannot fully explain why  a media outlet chooses to issue (or omit) a breaking news alert via email. Indeed one major news outlet, Yahoo, issues on average less than one breaking news email per week. This research will be able to shed light on whether or not media organizations with the same alleged bias cover breaking news in a similar way that reflects that bias.

1 comment:

Christopher Tunnard said...

There's a lot of potential here, and I hope someone takes you up on your offer to share the data. You are also right that whoever undertakes this should consult previous work on media bias, as what you recommend, seeing if bias affects breaking news, sounds like it might yield a too-predictable answer, and it's not obvious what SNA can bring to the mix. In fact, the more nuanced answer might not come from SNA, but from questions like how long into a breaking news story does it take (on average) for bias to appear, and does it vary by where on the spectrum the network is.