Growth/Article creation for new editors

From mediawiki.org

This is a potential project that would focus on improving onboarding and guardrails in the article creation process. This project is not currently prioritized, but is being considered as part of Growth's 2023-2024 Annual Plan.

Growth tools currently encourage newcomers to try small edits before diving into article creation, yet many newcomers still attempt to create an article right away. If this project is prioritized, we will work on ways to better onboard newcomers that register with an intention to create an article. This project would aim to provide new editors with better guidance and guardrails in the article creation process, with the intention of also lightening the load of new page reviewers.

Current status[edit]

User story[edit]

As a new editor interested in article creation, I want the proactive guidance and support I need to be successful, because I'll be more likely to continue editing if my initial contributions aren't reverted or deleted.

Who is this project meant to serve? Good faith newcomers on Wikipedia who are primarily interested in article creation. About 28% of newly registered users who complete the Welcome Survey indicate that they have created an account specifically to create a new article. This is the population we hope to initially focus on.

There are many types of newcomers who come to the wikis, some looking to make good faith edits and others who simply want to vandalize, spam, or make bad faith edits. However, this good faith / bad faith distinction is perhaps too simplistic as there are also many newcomers who are want to make "good faith" edits, but don't understand core Wikipedia principles and guidelines around notability, conflict of interest, neutral point of view, etc. These newcomers need additional guidance or they end up frustrated and disappointed when their edits or articles get reverted or deleted. Because they aren't receiving the proactive guidance they need, they end up creating additional work for content moderators who need to provide reactive guidance which may not always be well-received.

Previous research identified six new editor personas. Below is an exploration into how this project might impact these personas:

New editor personas How could this project help this new editor persona?
Social changer This project could help this persona better understand policies and guidelines like verifiability, notability, neutral point of view, so they can share their knowledge about a cause that interests them in a way that is appropriate to Wikipedia.
Knowledge sharer This project could help this user better understand the rules and policies on how content should be written and cited.
Reactive corrector This project would likely have limited impact on this persona, as "reactive correctors" make quick edits and correct obvious inaccuracies or mistakes and have less interest in article creation.
Audience builder This project could help this persona experience less frustration by surfacing the importance of citing reliable sources early on and provide up-front recommendations around conflict of interest (COI) editing.
Box checker This project could help make writing an article easier and more intuitive for newcomers.
Joiner-inner This project could help make article writing more accessible to someone who appreciates more guidance and has interest in learning a new skill and joining the movement, but is unsure where to start.

The Growth team works on tools that are meant for new editors, with the objective to improve newcomer activation and retention. However, it's important to recognize that edits made by newcomers are then reviewed by experienced editors. So we will also need to consider the patrolling burden associated with newcomer edits, especially around new article creation. The following experienced editor perspective should therefore be considered, although these use cases may represent work needed outside of the scope of this project:

As an experienced editor who interacts with newcomers during the article creation process,

  • I want tools that help quickly identify submissions appropriate for the encyclopedia
  • I want tools to help quickly assist newcomers when their submissions require additional work
  • I want tools to quickly process submissions that are not suitable for Wikipedia in a way that is newcomer-friendly
  • I want access to better metrics to help quantify the work done by myself and other patrollers because I want to use my time and expertise efficiently and effectively.

Goals[edit]

We want to:

  • Make sure good faith newcomers with appropriate article ideas can successfully publish their article
  • Provide guidance so that good faith newcomers are more successful when creating articles
  • Retain good faith newcomers and provide a good experience, even if their article idea isn't notable or appropriate for Wikipedia
  • Minimize the number of articles and drafts created that are not notable or not suitable for Wikipedia to help reduce content moderation burden

We don't want to:

  • Restrict article creation further or change user access levels
  • Create a process that limits bad faith article additions, but unintentionally also limits good faith article additions or reduces newcomer retention

Community discussion[edit]

Inclusive product development framework phases

The Growth team is guided by the Inclusive product development framework, and is committed to gathering feedback from a wide range of editors and wikis before moving this project forward. There are many different approaches we could consider to help address the underlying needs around article creation, so we are first gathering feedback from experienced editors, and reviewing existing ideas that have been proposed. After this period of discovery we will start to define the project scope have an additional community discussion in which we ask for feedback regarding a more well-defined project proposal.

Initial ideas (Strategize & Discover phase)[edit]

Many different ideas have been suggested by community members:

The Wikimedia Foundation's Editing team is also working on a project that has some potential overlap with this project:

Early high-level community discussion (Strategize & Discover phase)[edit]

An initial community conversation was started at ar, bn, cs, es, fr and en Wikipedia. We gathered high-level feedback, as we are still early in the strategize and discover phase of this project.

Overall, this project idea received considerable attention and feedback on larger wikis.  It's clear that reviewing articles from newer editors is a task that is especially challenging for larger wikis, and there are many competing ideas for how to more effectively manage this.  Although better guidance in the article creation process may help improve the quality of a new editor’s articles, it would be better to encourage users to improve existing articles first.

Some new editors are only interested in creating articles with a promotional purpose, which is very time consuming for established users. A new article creation system shouldn’t help them to achieve their promotional goals.

The quality of new articles is a real problem, not only because of promotional contents, but also because it is an overwhelming process. Newcomers often struggle with unclear guidance, missing sources, citing unreliable sources, or forget to include citations entirely. As a result, some well-intentioned new editors become discouraged. For moderators, it takes a lot of time to review edge-case notability pages, or to ask for better sources. Increasing the workload of reviewers is not a viable option.

Several possible solutions have been suggested:

  • Decision tree, to help determine if the article should be included in Wikipedia, or if the user is allowed to create one (case of conflicts of interest).
  • Quality gates, to prevent articles with too few citations, promotional wording, etc.
  • Highlight deletion data, to remind users that Wikipedia seeks high-quality content.
  • In-context guidance, with article templates or highlighting examples of existing articles.

All this process should be customizable by each community, to fulfill their needs.

Using drafts more is suggested by many users (on larger wikis). English Wikipedia has made article creation though draft mandatory for new editors On English Wikipedia, the ability to create articles directly in mainspace is restricted to autoconfirmed users, though non-confirmed users and non-registered users can submit a proposed article through the Articles for Creation process.

It's clear we won’t totally automate this process: humans (patrollers, new page reviews, mentors, etc.) play a critical role in engaging newcomers and providing feedback about new articles. Individuals have various perspectives as to how involved experienced editors should be in new article review; some ideas were mentioned around creating a more collaborative draft-writing process, or even a review of sources prior to the article draft review. But whatever feature we consider should ensure we aren't over-burdening experienced editors, and allow for communities to customize the feature to work for their unique needs. It is important to keep the wikis as a collaborative place, where humans interact with each other, and to avoid a bureaucratic effect that would repel newcomers from participating.  

The suggestions and insights shared by the community have been invaluable in identifying the challenges surrounding article review and new editor engagement. The Growth team will carefully consider all the feedback and will incorporate it into a refined project proposal for further community review soon.

Project proposal community discussion (Discover & Define Phase)[edit]

  • We will start an additional community discussion once we have a more well-defined project proposal to discuss.

Research[edit]

Previous research[edit]

One key research finding of the New editor experiences project, was that “New editors' greatest challenges are not technical but conceptual. They struggle to learn Wikipedia's policies and how to shape content "the Wikipedia way".

Article creation on Wikipedia:[edit]

Data related to article creation varies considerably across different language editions of Wikipedia based on various factors. Some major factors that seem to influence article creation include: the size of a wiki, restrictions based on user access levels, and the use of Flagged Revisions.

  • Across all Wikipedias from November 1, 2021 to October 31, 2022, about 4.26% of new articles created, are created by accounts that are less than 30 days old.
  • About 28% of newly registered users who complete the Welcome Survey indicate that they have created an account specifically to create a new article.
Article deletion on Wikipedia:[edit]

In 2022, across all languages and devices:

  • 12% of all mainspace articles created were ultimately deleted [1].

That percentage is higher when you narrow by new editors (as defined by accounts that are less than a week old). In 2022, across all languages and devices:

  • 43% of mainspace articles created by new editors were deleted. [2]

These metrics vary considerably per wiki, where different article creation flows and norms exist. On English Wikipedia for example, article creation is restricted to user accounts that are at least 4 days old and have made at least 10 edits, while Content translation on English Wikipedia is restricted to extended confirmed users. In 2022, on only English Wikipedia, across all devices:

  • 37% of mainspace articles created by newcomers were deleted. [4]

Research to guide this project[edit]

The Growth and Apps teams are looking to address moderator concerns around new articles, and understand and improve the experience newcomers have attempting new article creation. Potential qualitative research may include a literature review including a review of existing policies, interviews with moderators/patrollers, and interviews with newcomers.

Several pertinent quantitative questions that we hope to answer:

  1. For newly registered users who explicitly state their intention of creating an account for the sole purpose of generating a new article:
    • What proportion of these users actually go on to create an article within 30 days?
    • What percentage of the articles produced by this group end up being deleted?
  2. How do these statistics compare with newcomers who select a different reason for creating an account?
  3. In examining the long-term productivity, user retention, and revert rate, how do these aspects differ among various Welcome Survey groups? For example, we can compare users who signed up with the intent to "edit Wikipedia" with those who signed up to "create a new Wikipedia article" and those who joined with the aim of "fixing a typo."
  4. Focusing on the articles created by newcomers that are subsequently deleted, what are the primary reasons leading to these deletions? What is the distribution of deletions related to conflict of interest as opposed to issues of notability or the absence of references?

By addressing these questions through both qualitative and quantitative research, we aim to uncover key insights that guide this project.

Design[edit]

  • TBD

Measurement[edit]

  • TBD