Reports & Briefs

Platforms & Publishers

Reports & Briefs

Tow/Knight Projects

Who Retweets Whom: How Digital And Legacy Journalists Interact on Twitter

Project Leader:

When bloggers and citizen journalists became fixtures of the U.S. media environment, traditional print journalists responded with a critique, as this latest Tow Center brief says. According to mainstream reporters, the interlopers were “unprofessional, unethical, and overly dependent on the very mainstream media they criticized. In a 2013 poll of journalists, 51 percent agreed that citizen journalism is not real journalism”.

However, the digital media environment, a space for easy interaction has provided opportunities for journalists of all stripes to vault the barriers between legacy and digital sectors; if not collaborating, then perhaps communicating at least.

This brief by three PhD candidates at The University of Washington, Michael L. Barthel, Ruth Moon and William Mari, takes a snapshot of how fifteen political journalists from BuzzFeed, Politico and The New York Times, interact (representing digital, hybrid and legacy outlets respectively). The researchers place those interactions in the context of reporters’ longstanding traditions of gossip, goading, collaboration and competition.

They found tribalism, pronounced most strongly in the legacy outlet, but present across each grouping. They found hierarchy and status-boosting. But those phenomena were not absolute; there were also instances of co-operation, sharing and mutual benefit. None-the-less, by these indicators at least; there was a clear pecking order: Digital and hybrid organizations’ journalists paid “more attention to traditional than digital publications”.

A note on methodology

The Tow Center posted this paper on March 5th 2015, since then, we have found a need to address questions about the conclusions, methodology and discussion, which were exacerbated by some factual errors (that have now been transparently corrected).

The Tow Center publishes work in a number of different forms from different types of researchers; working journalists as well as academics. This introduces a diversity of subject matters, of research approaches, and of norms. This piece presented preliminary research, as part of a series on social media, journalism and citizenship organized by Professor Phil Howard and a team of researchers at the University of Washington. The text of the study said that the results were preliminary, that the sample size was relatively small (750 tweets from 15 selected journalists) and collected on a single night in October 2013.

In the paper, the authors Michael Barthel, Ruth Moon and William Mari explained the process of going from research questions to a pilot study:

“To start answering these questions, we undertook a preliminary analysis of the Twitter use of journalists from these three types of organizations. Our analysis is based on a human examination of a moderate number of tweets as opposed to a computational analysis of a large data set. We favored the deeper holistic view this manual approach brings over a wider automated process. Later studies may be able to convert our process into an algorithmic approach and examine a bigger set of data.”

In many examples of preliminary research, sample sizes are small, and conclusions may be contested.

Professor Howard continues:

But research often involves purposefully sampling from evidence so as to develop some hypotheses. The findings from this preliminary analysis can now become the testable hypotheses behind further qualitative, comparative, or quantitative research. It is difficult to generate new knowledge without working hypotheses, because these are needed before anyone else can confirm, modify, or reject what has been revealed here.

The Tow Center always welcomes feedback, positive and negative, on every piece of work we publish, as it helps us refine our research and be of more use to the journalist community.

The full brief is fascinating reading: You can download your copy here (pdf).


We have made a number of corrections to this paper. 
On page 11, paragraph 1, the incorrect name spelling “Nicolas” Confessore has been replaced with Nicholas Confessore. 
Also on page 11, paragraph 1, the paper said “the five Times reporters admitted to our sample were Jonathan Martin (a different Jonathan Martin works for Politico)”. That was incorrect. Jonathan Martin moved from Politico to The New York Times. The words “(a different Jonathan Martin works for Politico)” have been removed.
On page 14, paragraph 1, the paper had “the Times’ most-followed political reporter, Jonathan Martin of Fox News”, which seemed to say that Mr Martin worked both for the Times and Fox News. Mr Martin did not work for Fox. The words “of Fox News” have been removed.
On page 11, paragraph 1, the words “The five Times reporters admitted to our sample” have been replaced with “The five Times journalists admitted to our sample” to reflect that some held more senior roles than reporters.
On page 11, paragraph 3, the words “We consulted the Politico masthead for a list of its reporters” have been replaced with “We consulted the Politico masthead for a list of its editorial staff” to reflect the fact that it included editors and other senior editorial staff.


Philip N. Howard is a professor in the Department of Communication at the University of Washington, with courtesy appointments in the Jackson School of International Studies and the Information School. Currently, he also works as the Director of the Center for Media, Data and Society and the founding Professor of a new School of Public Policy for Central European University.

March 5, 2015