liveBook · Manning

Nudge

As explained in Hidden Influences

A user-interface intervention that prompts a person to reconsider an intended action without directly preventing it.

Before a heated post is shared, a platform might pause and warn that similar content was previously removed or could be offensive. The user can still continue, but the extra moment invites reconsideration. A nudge is this kind of interface intervention: it prompts a person to reconsider an intended action without directly preventing it. section 3.4.1

A platform nudge can work by:

  • Scanning content before posting with a machine-learning filter. section 3.4.1
  • Displaying an explanation that similar content was actioned or might be offensive. section 3.4.1
  • Creating an extra step that gives the user an opportunity to revise or delete the content. section 3.4.1
  • Using social information, such as friends’ actions, to make a desired behavior more salient. section 7.2

Two platform experiments

Twitter reported fewer shared harmful posts after warning users before publication, partly because users revised or deleted them. Facebook’s 2010 voting experiment compared a reminder with friends’ photos, a reminder without those photos, and no change; the socially enriched reminder increased clicks on the “I Voted” button. Clicking, however, measured online signaling rather than voting itself, so researchers also compared profile names with public voting records. section 3.4.1 section 7.2

Nudges are not neutral simply because they leave the final click to the user. A false positive can wrongly label harmless speech, and warnings around political or controversial topics may suggest that a platform is pursuing an agenda. The intervention therefore combines behavioral influence with the risks of opaque and imperfect algorithmic decisions. section 3.4.1 section 7.2

How Hidden Influences covers it

How deep it goes
Explains scanning, warnings, extra steps, social cues, and measurement limits, while stopping short of implementation details.
Examples
Uses pre-posting warnings, Twitter’s harmful-content experiment, and Facebook’s voting reminders with or without friends’ photos.
What you can do after
Evaluate nudges by distinguishing changed clicks or signals from changed behavior and by examining false-positive and agenda risks.
How it connects
Connects nudges with content moderation, machine learning, algorithmic decisions, social influence, and political mobilization.

Also explained in

Compare how 2 books explain nudge →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Hidden Influences. The section links point to the parts of the book it draws on.