9483
AiS
Session 8b
Joshua
Foster
Simulation instructions.
Simulation instructions.
To summarize:
What happened in the simulation?
What specifically made this task difficult?
Rational herding.
In a social learning environment, it is often rational for an individual to ignore their private information and follow the "herd" (i.e., adopt the beliefs and actions of those they observe).
Mapping the Simulation to Social Feeds
| Simulation | Social Media |
|---|---|
| Private draw | Personal experience / niche facts |
| Public guess | Like/reshare/comment counts |
| Order of moves | Ranking + timing on the feed |
| Payoff | Engagement, identity rewards, reach |
| Herding threshold | Trust in sources; scarcity of attention |
Why are network effects so important to social media companies?
Reed's Law.
Social networks increase in "value" in proportion to $2^n$ of their size $n$, which allows for the formation of subgroups, as they allow for an enormous number of potential group formations.
How many connections to you have on LinkedIn?
How many have you communicated with in the last 3 months?
Dunbar's Number.
An upper-bound estimate on the number of relationships an individual can maintain at once.
How do Reed's Law and Dunbar's Number jointly shape the spread and evaluation of true and false information on Facebook?
Hint: think about information volume vs. attention and group scale vs. trust capacity.
What behavioural biases may be playing a role in the proliferation of "fake news" on social media platforms?
Motivated Reasoning.
A cognitive process in which justification is provided for the individual's preferred outcome, despite the presence of contradictory evidence.
Numeracy Score
What would the objective of a welfare maximizing policy be with respect to "fake news"?
What responsibility does Facebook have in implementing this policy?
What does the case suggest Facebook can do regarding political speech?
Which option do you recommend for Facebook?
Does Facebook have sufficient incentive to implement a welfare-maximizing strategy against "fake news"?
Is there a public solution that is better?
What might a regulator do here?
Key Takeaways.
Incorrect information cascades can be the result of:
Either can lead groups to incorrect beliefs despite many individuals being presented with correct information.