According to the passage, heavier users of social media encounter more opposing views than lighter users because

platforms deliberately insert opposing content into feeds.

such users actively seek out sources they disagree with.

opposing content generates more engagement than agreeable content.

a feed carries whatever loose acquaintances post, which is broader than what people choose offline.

The observation about which age groups have polarized fastest is offered as evidence that

the rise in polarization does not track the exposure the theory says causes it.

older people are more susceptible to persuasion than younger people.

polarization has been measured inconsistently across countries.

television remains a more powerful influence than the internet.

The statement that "exposure is not persuasion" serves chiefly to indicate that

people rarely remember content encountered incidentally in a feed.

the studies measuring exposure used unreliable methods.

a diverse feed may harden positions rather than soften them.

opposing views are usually presented accurately on social platforms.

The passage indicates that the sorting of people into like-minded groups is

largely a product of the networks people build themselves, which platforms amplify.

produced by ranking algorithms and would disappear without them.

confined to countries with highly partisan news media.

a recent phenomenon with no offline precedent.

Which of the following, if true, would most strengthen the original filter bubble theory as described in the first paragraph?

Users report spending more time on platforms than they did five years ago.

Journalists and legislators continue to cite the theory frequently.

Platforms have begun publishing details of how their ranking systems work.

When ranking is switched off and feeds are shown in chronological order, users encounter substantially more opposing content.

←

The Filter Bubble and the Evidence Against It

PRACTICE

100:00+4 −1MCQ Single Answer

Reading Passage

The filter bubble is among the most successful social theories of the past two decades. Its claim is straightforward: algorithms rank content by predicted engagement, engagement is highest for material a user already agrees with, and the result is an information environment that quietly walls each person inside their existing beliefs. Fragmented publics and hardening partisanship follow. The idea has been adopted by journalists, legislators and regulators, and it now shapes how platforms are written about and governed. The empirical record is far less obliging than the theory's popularity suggests. Studies that measure the range of political content people actually encounter have repeatedly found that heavier users of social media and search meet more views opposed to their own than lighter users do, not fewer. The mechanism is unglamorous: a feed delivers whatever a person's loose acquaintances happen to post, and that incidental exposure is broader than the deliberate choices most people make offline, where friends, neighbourhoods and newspapers are selected with more care. The timing is awkward for the theory as well, since in several countries polarisation has risen fastest among the age groups least likely to be online at all. None of this shows that platforms are harmless, and the more careful critics have moved to firmer ground. Exposure is not persuasion: encountering an opposing argument framed as an attack can entrench a position rather than soften it, so a diverse feed may polarise as readily as a narrow one. Sorting also happens well upstream of any algorithm, in the networks people build for themselves, which platforms amplify rather than invent. The reframing that survives is narrower than the original and more useful. The problem is less that people see too little disagreement than that most of the disagreement they see arrives stripped of the context that would make it intelligible.

The main point of the passage is that the filter bubble theory