Wikipedia:Article Feedback Tool
|Type||Assessment (wiki content)|
|Website||Article Feedback @ Mediawiki
(user help: Wikipedia:Article Feedback Tool)
- Note: Version 5 of the Article Feedback Tool is now under development by the Wikimedia Foundation. To hear more and give your feedback on this new feature, please go to the AFT V5 project page.
This tool was created with the following goals:
- Quality assessment – Article feedback complements internal quality assessment of Wikipedia articles with a new source of data on quality, highlighting content that is of very high or very low quality, and measuring change over time.
- Reader engagement – Article feedback encourages participation from readers, offering a call to action for some assessors to improve the article.
The current version of the tool features a rating form at the bottom of each Wikipedia article, inviting readers to rate that article based on four criteria. This version (V4) is currently deployed on almost all pages on the English Wikipedia, among other wikis.
A new Version 5 of the Article Feedback Tool (AFT V5) is now under development, based on community feedback and analysis of the public data for the current version. The main goal of this version is to find ways for readers to contribute productively to building the encyclopedia.
The goals of the current version of the tool are to 1) provide a mechanism by which readers can help assess the quality of an article and 2) provide an on-ramp for readers to start engaging with Wikipedia. The current tool invites readers to rate this page for these four qualities: trustworthiness, objectivity, completeness and writing (see image).
"Note: Registered users may disable this feature by going to Preferences→Appearance, and checking the "<articlefeedback-disable-preference> " box."
WMF is working on the next version (Version 5) of the tool. Analysis of the current version and feedback from the community are being incorporated into the design of the next version. The following blog post describes the features of the tool that appear to be working versus features that are not quite as successful at quality assessment and reader engagement.
The next version of the tool will be focused on finding ways for readers to contribute productively to building the encyclopedia. When Wikipedia editors were asked over IRC to comment on what types of feedback would be useful to hear from readers, some mentioned that it would be useful to know what readers were looking for and if they found it. One direction that will be tested in the next version is a simplified form which asks the reader "Did you find what you were looking for?", with 'Yes / No' buttons. Readers will also be invited to add a comment or suggestion for improving the article.
Quality is still an important consideration and the next version of the tool will continue to test various ways of measuring quality. In a departure from the current tool, we may test implicit measurements of quality. For example, with the "Did you find what you're looking for" question, the percentage of "yes"'s could be an indicator of article quality.
The first phase of the feedback project began in 2010, with the started on 9 July, the Workgroup formed on September 15 and a project page created on October 7. Various functional requirements were developed in conjunction with English Wikipedia editors via discussion (e.g., at Village Pumps) and announcements (in The Signpost and at Administrators' Noticeboards).
Phase 2 began on TBD, and a wikicategory for the project started on February 14, 2011. In July 2011, "What's this?" help function on the "Rate this page" box replaced the phase 2 "Please take a moment to rate this page" wording.
Surveys that thousands of users took in the summer[when?] of 2011 show that more than 90% of users believed the tool was useful, and slightly more than half believed that it positively affected the development of articles.
In the fall[when?] of 2011, users typically submitted between 30,000 and 45,000 separate ratings each day. About 97% of users submitting ratings were not logged-in editors, indicating significant uptake by readers.