,
12 tweets,
3 min read
Read on Twitter
New newsletter discussing insights from game theory on political spin.
tinyletter.com/hiddenincentiv…
Summary in this thread. Based on ongoing work with @moshe_hoffman 1/12
tinyletter.com/hiddenincentiv…
Summary in this thread. Based on ongoing work with @moshe_hoffman 1/12
What are common features of spin? Why are they so common?
I'll illustrate with examples from recent political "debates" on social and mainstream media. Though such spin shows up in many other contexts (e.g., dating profiles, scientific research). 2/12
I'll illustrate with examples from recent political "debates" on social and mainstream media. Though such spin shows up in many other contexts (e.g., dating profiles, scientific research). 2/12
1. People only reveal supportive evidence; they obscure evidence that is unsupportive.
Old enough to remember this one?
3/12
Old enough to remember this one?
3/12
Or this one?
4/12
4/12
2. People only search hard for supportive evidence; they do not search for evidence that is unsupportive.
An example of searching for evidence: watching the news (we agree with). Can provide us with examples to support our political positions.
newrepublic.com/article/119922… 5/12
An example of searching for evidence: watching the news (we agree with). Can provide us with examples to support our political positions.
newrepublic.com/article/119922… 5/12
3. People present putatively supportive evidence
E.g., by cherry picking facts
climaterealityproject.org/blog/three-way… 6/12
E.g., by cherry picking facts
climaterealityproject.org/blog/three-way… 6/12
For each of these types of spin, the key puzzle: these behaviors persists even though everyone knows people are engaged in them (and discount what they hear accordingly). 7/12
To try to address this puzzle, we:
1. presume that spin is driven by persuasion motives
2. develop simple models of such persuasion
3. show that the above features of spin are part and parcel of the models' equilibria. 8/12
1. presume that spin is driven by persuasion motives
2. develop simple models of such persuasion
3. show that the above features of spin are part and parcel of the models' equilibria. 8/12
The key assumption: private information. Only we know:
- whether unsupportive evidence was obtained but withheld?
- how intensive was search?
- which “tests” were run?
9/12
- whether unsupportive evidence was obtained but withheld?
- how intensive was search?
- which “tests” were run?
9/12
Key result: We don’t reveal private information when it harms us, even if others know we are concealing, and adjust their beliefs accordingly.
I.e. spin is an optimal strategy even though we all know people are engaged in it, and discount for it. 10/12
I.e. spin is an optimal strategy even though we all know people are engaged in it, and discount for it. 10/12
Of course, these features of spin are only expected from those who are motivated to persuade. If they are motivated to inform or discover the truth, will reveal and search for evidence in unbiased fashion. 11/12
So, when we see our three features of spin:
- Only present supportive evidence
- Only search for supportive evidence
- Present putatively supportive evidence
We can be sure a media outlet, politician, or advisor’s goal is to persuade, not to inform. 12/12
- Only present supportive evidence
- Only search for supportive evidence
- Present putatively supportive evidence
We can be sure a media outlet, politician, or advisor’s goal is to persuade, not to inform. 12/12
