January 20, 2016

Crowdfunded Journalism: A Small but Growing Addition to Publicly Driven Journalism

Methodology

This report examined funded projects in Kickstarter’s Journalism category between April 28, 2009 and Sept. 15, 2015. The aim of the research is to gain a better understanding of the status and dynamics of crowdfunding in the context of journalism. Kickstarter, recently reincorporated as a public benefit corporation, is one of two larger-scale and broad-based crowdfunding platforms, but currently the only one with a distinct offering for journalism projects. The other, Indiegogo, launched in 2008, offers categories for film, photography, transmedia and video/web projects – but does not have a specific journalism genre. Another effort, Beacon, launched in 2013 as a platform dedicated to crowdfunding journalism based on a subscription model but now focuses more on partnerships and matching funds. Several other narrowly targeted crowdfunding platforms dedicated to journalism entered –and left – the space within the past few years, including Spot.us for community-funded journalism, Contributoria and Indie Voices for independent media, Emphas.is for photojournalism and Vourno for video journalism. An analysis of Kickstarter, then, while not completely exhaustive of all possible journalism-related crowdfunding, seemed an appropriate way to gauge the status and dynamics of this emerging arena.

For this project, researchers analyzed publicly available data for the 2,975 proposed projects in Kickstarter’s Journalism category between April 28, 2009 and Sept. 15, 2015, with an in-depth examination of the 669 journalism projects that received full – or more than full – funding.9 (For more on the manual coding of this study click here.)

Data gathering proceeded in two stages. In the first, researchers used automated tools to obtain a list of all projects in the Journalism category and to extract basic data about each project. In the second stage, the 669 funded projects were manually coded for a number of specific elements not broken down in the publicly available data, including the type of producer, what specific types of journalistic endeavors were being presented to potential backers, the geographical orientation and whether a project was a new initiative or an expansion of ongoing work.

It is worth noting that when Kickstarter first launched as a crowdfunding platform in April 2009, there was no dedicated category for Journalism. The company made the decision in mid-June 2014 to introduce two new categories: Journalism and Crafts. At that time, the company also retroactively recategorized earlier, relevant projects into the Journalism category, and created five new Journalism subcategories going forward: Audio, Photo, Print, Video and Web.

Researchers reached out to Kickstarter staff on several occasions during the early stages of this project to obtain more knowledge about the specific content of the Journalism and other categories, how the categories were assigned, the format and structure of the data on the website and the basic methodological approach of our research. All research decisions and editorial input were made in-house.

Automated data acquisition

On Sept. 15, 2015, researchers accessed Kickstarter’s page listing all projects in the “Journalism” category (https://www.kickstarter.com/discover/categories/journalism).

Using an automated process that can extract information from Web pages, researchers extracted the links to all 2,975 projects classified in the Journalism category at that time. Researchers then used a different automated process to visit each project’s page, download the HTML code, and automatically extract some information:

  • Project status:
    • Funded (reached end date and achieved goal)
    • Unfunded (reached end date but did not achieve goal)
    • Cancelled (withdrawn by creator before reaching end date)
    • In process (not yet reached end date)
  • Number of backers
  • Amount pledged
  • Goal amount

Among these 2,975 projects, 1,941 were classified as unfunded, 246 as cancelled, 119 as in process (“live”) and 669 as funded. For this analysis, researchers focused on the 669 Journalism projects with a funded status.

Manual coding of Kickstarter projects

Two experienced researchers, specifically trained for this project, manually coded the 669 funded projects by considering the information contained in all but one section of the individual project profiles – including “Created by,” the entirety of text descriptions of the projects, and videos, when present. (The one category not included in the coding analysis was the Rewards section.) The unit of analysis for this study was the individual Kickstarter project.

The six variables used in this study, briefly described, were:

  • Year of launch
  • Geographical region – Denotes where the project was to be conducted. Researchers coded for one of three options:
    • U.S.
    • Non-U.S
    • U.S./Non-U.S. combination
  • Travel – Denotes whether or not the project involved travel beyond a single metro area or single region.
  • Producer type – Researchers coded for one of four options:
    • Single individual: A single individual with no stated affiliation with any media outlet or organization, or other institution and acting on his or her own behalf.
    • Group of individuals: A group of two or more individuals, with no stated affiliation with any media outlet or organization, or other institution and acting on their own behalf.
    • Media outlet or organization: Any kind of outlet or organization whose primary mission is to produce news or information, and was in existence prior to the proposal’s launch on the Kickstarter platform. When an outlet or organization was named as the project creator but did not, in fact, yet exist, researchers defaulted to single individual or group of individuals, as appropriate.
    • Institution: Any kind of public or private institution other than a media-focused one. Included schools at any level (elementary through university), nonprofits, and for-profit businesses.
  • Primary product – Describes the primary kind of journalism being produced by the project. Researchers coded for one of 16 options:
    • App
    • Audiobook
    • Blog
    • Book
    • Documentary
    • Exhibit
    • Magazine
    • Multiple format
    • Newsletter
    • Newspaper
    • Photojournalism
    • Radio/podcast
    • Research paper
    • Website
    • Zine
    • Other/Not format-specific
  • Startup or Expansion – Designated whether the project was a wholly new initiative, or if it existed in any tangible form prior to the launch of the proposal on Kickstarter. Examples of expansion projects would include one for a print edition of something that already existed online, or funds to print a book when the research or reporting had already been completed.

Upon the completion of the coding of the 669 projects, the data cleaning process identified 11 projects among those that had been coded as “other” that contained no mention of journalism or the intent to produce a journalism product, and these 11 projects were removed from the dataset prior to the analysis, resulting in a dataset of 658 projects.

The vast majority of the projects (84%) requested funding in U.S. dollars (554 of 658 projects). For this analysis, funding amounts requested in other currencies were translated to U.S. dollars using conversion rates as of Sept. 15, 2015, on Oanda.com. Researchers also analyzed the total amount of funds both requested and pledged in foreign currencies using OECD average monthly conversion rates as of the month that the projects were launched on Kickstarter. The analysis revealed that the difference between the two methods produced only a 1% difference in the total overall level of funding across all projects.

Intercoder testing

Two experienced researchers worked together to learn and refine the codebook.

In order to demonstrate the validity of the coding rules that were specific for this project, intercoder testing was conducted on all of the six variables. Both researchers coded two sets of 10 projects and one set of 50 projects during the training process, and then completed coding a final intercoder set of 70 projects (slightly more than 10% of the total dataset) prior to conducting the actual coding.

The coding results for the set of 70 projects were assessed using Cohen’s Kappa.

The intercoder results according to Cohen’s Kappa were:

  • Year of launch: 1
  • Geographical location: .895
  • Travel: .909
  • Project creator: .916
  • Primary product: .933
  • Startup or expansion: .827
  1. During the data cleaning process, 11 projects were removed, resulting in 658 projects for analysis.