In my ECONS102 class last week, we covered the economics of information. One aspect of that topic is to look at media bias. So, I was interested to read this post by Nate Silver from last month (which I only caught up to read this week). Silver notes at the start of the post that his core hypothesis is:
In American media and political discourse, there has been a fundamental asymmetry during the Trump Era. Left-progressives, liberals... centrists, and moderate or non-MAGA conservatives all share a common argumentative space. I call this space the Indigo Blob, because it’s somewhere between left-wing (blue) and centrist (purple). The space largely excludes MAGA/right-wing conservatives — around 30 percent... of the country.
The rest of the post is about the causes and consequences of the 'Indigo blob', and I encourage you to read the whole post. However, I want to focus in on this bit:
This graphic from a paper by Magdelena Wojcieszak, et. al. reflects the ideological position of different groups of Twitter users from 2016-2019. There’s a lot going on, so let me show you the figure first and then we can talk our way through it.
Among politicians on Twitter, we see a traditional bimodal distribution with liberal and conservative peaks corresponding to the Democratic and Republican parties. But everything else is left-skewed. There’s not a clear distinction between left-wing and centrist news organizations — they’re all part of the Indigo Blob — and the right-wing media isn’t terribly well-represented. The same holds for journalists, although the right flank is even more attenuated. And Twitter’s users during this period skewed heavily left too, with liberals outnumbering conservatives roughly 3:1, according to Wojcieszak. Anyone who was active on the platform during this period knew that content satisfying to progressives and Democrats was more likely to be rewarded with likes and retweets.
I believe this contributed to the misestimation of the political preferences of the American public by journalists, Twitter-savvy politicians and others who were active on the platform.
I want to co-opt that figure to go in a slightly different direction, because it highlights two key facts about bias among journalists. These two key facts are well explained by some simple models of media bias. I'm not going to explain those models here, but they are described in this article I published earlier this year in the Journal of Economic Education (ungated earlier version here).
First, notice that the mass of journalists is mostly to the left of the centre. There is an overall bias in the media towards the liberal side of the spectrum (in that there are more media outlets and more journalists on the liberal blue side than on the conservative red side). However, notice that this follows the overall profile of Twitter users, albeit imperfectly. There are more Twitter users on the liberal blue side than on the conservative red side. So, to some extent it is understandable that journalists and media would be biased in this way. It reflects the characteristics of demand from the consumers of media. If there are more liberal media consumers, then there will be more liberal media suppliers.
Second, there is a distribution of media across the spectrum. Some media outlets and journalists lean liberal, and others lean conservative. Collectively, media outlets and journalists cover the range of views, but each media outlet (and to a lesser extent, each journalist) targets a relatively narrow audience within the spectrum. They don't tend to appeal to media consumers with views that are far from their own, because media consumers tend to prefer to consume news that accords with their own views - a behavioural economics concept known as confirmation bias.
The takeaway message from those two facts is that the media are biased. This isn't necessarily something to fear, but it is something to acknowledge and recognise. Possibly we should worry a bit more about the bias in the politicians? [*]
[HT: Marginal Revolution]
*****
[*] Of course, we should note that the population of Twitter users will not necessarily be representative of the population overall. So, perhaps media are biased relative to the population as a whole, and politicians are less biased than the media?
No comments:
Post a Comment