Thanks so much for sharing this profile of my work, @econuoft.bsky.social! It was a pleasure chatting about my research, teaching and involvement in the academic community.
In the experiments, I frame the game as follows:
- One "plant pot" contains a seed that will grow more if a sprinkler closer to it is turned on
- A Chooser decides whether to accept a recommendation to switch sprinklers, but is wary of the recommendation to switch to sprinkler 3 [3/7]
#EconSky #EconJMP
A Chooser should reject a recommendation to switch to sprinkler 3. Thus, an Advisor’s optimal recommendations involve recommending sprinkler 2 when pot 3 is the plant pot. Intuitively, this prompts the Chooser to make a small welfare-improving change, instead of none at all! [4/7]
To study whether paternalistic “Advisors” can persuade reluctant “Choosers” to change their behavior, I:
1️⃣Develop a model of a recommendation game, which characterizes the recommendations an Advisor *should* send.
2️⃣Experimentally test whether Advisors *actually* send them. [2/7]
Could you persuade someone who's reluctant to change their behavior to take an action that’s in their best interest? I find that, on their own, most people can't! A🧵on my JMP [1/7]
🔗to paper: alexballyk.github.io/personal-web...
🔗to my website: sites.google.com/view/alexbal...
#EconSky #EconJMP
In my Main Experiment, I find:
- Despite learning that Choosers consistently reject a recommendation to switch to sprinkler 3, most Advisors send sub-optimal recommendations
- Many Advisors always recommend the sprinkler above the plant pot [5/7]
Alex Ballyk
My results imply that...
- Advisors may not consider how Choosers will respond to their recommendations
- Prompting them to do so may improve their ability to steer Choosers towards better decisions
Read more here: alexballyk.github.io/personal-web...
Thanks! Feedback is welcome!🙂 [7/7]
Alex Ballyk
In my Beliefs Experiment, I prompt Advisors to guess the Chooser’s strategy before sending their recommendations. I find that Advisors become much more likely to send optimal recommendations. [6/7]
