Here's the code used to run the above simulations.
Can this regulatory approach be improved?
Suppose that we were to make the following modification:
• Instead of the regulator saying "both extremes are wrong", the regulator could identify the magnetization of the echo chamber and specifically say "the extreme on the side our community favors is wrong."
• The allegiance value would only shift in the appropriate direction.
• Only the opinion of that corresponding extreme would be suppressed.
The result surprised me when I saw it.
That slight change of targeted statements of "this one extreme position is wrong" had the effect of drastically depolarizing the populations, and (as shown below) even demagnetizing (or sometimes, flipping the magnetization) of some of the chambers.
Just a moment, I found a bug in my code. I didn't correctly handle the extremist opinion suppression.
I think I fixed the bug in my code. Here's the corrected version of the position/allegiance graphs for the 12 simulations:
And the chamber-by chamber breakdown for simulation 1 (with an overall negative magnetization)
As well as chamber-by-chamber analysis for simulation 2 (with neutral magnetization)
