News Curation and Discovery

Hi Everyone! Here is a quick blog post to keep you updated on what’s happening with the Firefox Context Graph team.

Over the course of this year, we published various posts about our studies on content discovery on the web. At the Mozilla All-Hands this summer, we began to drill down and focus specifically on news discovery.

In this post, I will briefly discuss why we believe access to high-quality news is critical to the health of the Internet. I will also introduce a study currently underway on how people get and share news — and present some fpreliminary data and analysis.

Keep in mind: This blog post is as much about us sharing the data, as it is about maintaining a channel of open communication with you — enthusiastic Mozillians, Firefox users, and Context Graph study participants. Without further ado, please read on and let us know your thoughts!


Digital Journalism and (Failed) Promise of the Web

An early promise of the web is that the Internet is supposed to enable a free flow of information, open up communication, and — as far as journalism is concerned — revolutionize the news media. Bloggers, netizens, and any “citizen journalist” would be able to publish content online. Each one of us could become an active investigator instead of a passive consumer of news. Readers could communicate with publishers, interactively explore the latest news, and receive personalized updates.

However, even though the web is initially designed to be open and decentralized, our access to information is increasingly monopolized by a handful of companies. As a result, our online experience today is often one where we live an echo chamber amplified by our respective social medial filter bubbles. Pew Research’s latest survey on Americans’ attitudes about news media finds that only 5% of adults strongly trust information they get from social networking sites such as Facebook and Twitter.

Some refer to the current state as the failed promise of the web.

High-Quality News and Internet Health

We believe a better understanding of how people access news — and effective tools to help people discover high-quality news — are critical to the health of the Internet.

To clarify, we intend to focus on both hard news (e.g., political events that make the headlines of the New York Times) as well as soft news (e.g., weather, morning traffic, where to eat, what to do this weekend).

Give Control Back to Users (and Publishers)

As our access to information online becomes centralized, publishers — more than ever—increasingly depend on a few major publishing platforms to distribute their content.

These platforms frequently apply algorithmic means to filter and promote content. When issues arise, such proprietary algorithms cannot provide explanations, receive feedback, or be held accountable for their actions. Furthermore, shifting algorithms on these platforms disproportionately hurt independent media, who have fewer resources to respond to the constant changes than the bigger agglomerated news sites.

We plan to study the effects of algorithmic distribution of news, how they compare to traditional news curation, and how much they adhere to established journalistic standards.

For Firefox users, we hope our research will lead to tools that help you discover relevant and high-quality news — outside the control of major publishing platforms or ad networks.

Equally important, by lending support to small independent publishers, we hope to help restore the original vision of a decentralized and open web — where anyone has a voice.

Help You Determine and Find Trustworthy Information

Let’s face it: 5% trust is an awful and woeful number for information found on social media — or anywhere else for that matter. A canvassing of expert opinions about online trust by Pew Research finds that the public’s suspicion of the web and social media have coincided with a sharp decline in trust for major institutions, such as the government, the news media, public schools, the church, and banks.

We plan to examine how people engage with news online, and better understand the factors and activities that impact people’s trust. Some current issues include the massive networks of fake accounts used to send spam, sow distrust and confusion, and/or artificially boost the popularity of select news; as well as the organized and intentional spread of fake information on the web, some of which have led to real-life consequences in U.S. communities.

We hope our research will lead to tools that help deter or minimize the actions of such bad actors on the Internet — and help you determine and find trustworthy content on the web.


Keeping Up with, Getting, and Sharing News

As a starting point, we are currently conducting a study to understand how people engage with news online. The study builds on Pew Research’s earlier work on how people get information about their local communities.

The data below are from the first 8,400 respondents of our survey. I’ll present some preliminary findings from two of the questions:

  • Do you enjoy keeping up with the news?
  • How do you get and share news online?

Age and Location

As Internet usage is strongly correlated with age and somewhat related to location, we made the decision to ask for these two pieces of demographic information in this survey.

The chart below shows the estimated number of Internet users in the United States vs. the distribution of our respondents by age. We have an over-representation of users between ages 20 and 37, an under-representation across all other ages, and some potentially dirty data at ages 18 and 99. For the remainder of this blog post, we reweighted our data to match the U.S. population by age.

The chart below shows the distribution of our respondents by location. We have more users in large cities (compared to 22% as reported in the original Pew research), an equal proportion of suburbans (compared to 21%), and fewer users in small cities (37%) and rural areas (20%).

Enjoy Keeping Up with the News?

We asked our users whether they enjoy keeping up with the news, a question that was part of the original Pew research.

The chart below shows the results obtained by the Pew Research in 2011, broken down by location. They found that suburban residents enjoyed keeping up with the news the most.

The chart below shows our results broken down by location. While the absolute magnitudes of our responses are different, we find that the rank ordering of our results remains very close to the original Pew research — both at the boundary “a lot/some” and at the boundary “some/not much.”

As we analyze the results by age, we observe a gradual decline in interest — with each older generation reporting higher interest than the succeeding younger generation in keeping up with the news. (Pew Research did not release the breakdown of their data by age.)

Ways of Getting and Sharing News Online?

Pew Research Center polled their users about eight means of getting and sharing news online. We replicated the question using the same instructions but added an additional option (marked below by *).

We asked our users whether they have ever gotten or shared news by:

  • Emailing a news article
  • Commenting on a news article
  • Contributing to an online discussion
  • Tagging online news content
  • Contributing their own news articles

Or,

  • Posting news articles on any social media
  • Posting news articles on Facebook (*)
  • Posting news articles on Twitter

Or,

  • Customizing their Firefox/browser homepage

For all non-social media engagements, we again observe numbers that are different in absolute magnitudes but identical in rank ordering to the original Pew research. They include the five forms of online engagement (email, comment, discussion, tag, contribute) and one type of browser interaction (customize homepage).

On the other hand, we observe a decline in social media engagement — both across all social networking sites and on Twitter. Engagement on Facebook is almost as high as engagement across all social networking sites.

At this point, it is unclear whether the decline is due to the changes in the data collection method (web survey vs. phone interview), population differences (Firefox users vs. general U.S. population), a real temporal pattern from 2011 to 2017, or other causes.


Next Steps

In the coming weeks, we will continue to examine online news engagement. In keeping up with principles of transparency here at Mozilla, we will continue to keep you posted on what we are doing, how we are collecting your data (always with explicit consent), and using your data.

Of course, all this research is only possible with your participation and contribution. We would like to keep an open channel of communication, and welcome your comments and feedback at any time.