Personalisation narrows the information a person encounters to what already fits them.
Pariser's 2011 term. Empirical support is mixed — several studies find online news diets more diverse than offline ones — so the mechanism is plausible and the magnitude is disputed.
In practice
- A recommendation feed trained on three months of viewing offers a fourth month that looks much like the third.
- A reader who follows twelve accounts on one platform meets a narrower range of argument than the same reader gets from one newspaper's letters page.
- Possible explanations
- One explanation is algorithmic personalisation. Another is self-selection, since people choose congenial sources unprompted. A third is homophily: feeds narrow because friends are similar, before any algorithm acts.
- In its favour
- Personalisation does narrow what is shown in controlled tests, and sharing networks are measurably homophilous.
- Against it
- Survey and behavioural studies repeatedly find online news diets more diverse than offline ones, and large platform experiments have found small effects on attitudes, so the magnitude is disputed even where the mechanism is real.
- Standing
- Empirical support is mixed; several studies find online news diets more diverse than offline.
- Further reading
- Wikipedia
- Subject
- Communication & Media
- Essay
- Theories of what stands between an event and its audience