The Diffusion of Innocence

September 15, 2012 at 5:16 pm 1 comment

| Gabriel |

[Update 1: From skimming Al Jazeera’s (conveniently date-stamped) blog on this issue for Saturday and Sunday it looks like the protests have slowed considerably, which would imply an s-curve.]

[Update 2: It looks like the prime mover political entrepreneur was the shock artist, who actively tried to get a reaction out of people. That is, this is more similar to Jones threatening to burn Korans than to the Danish imans going on tour with (forged versions of) the Danish cartoons. Of course as in purely domestic culture wars issues, there can be a strange symbiosis between partisans on both sides who disagree on the merits but mutually benefit from discord.]

I took the Atlantic Wire’s map (also see the KML file) of “Innocence of Muslims” protest, did my best to add dates, and graphed it as a (cumulative) diffusion curve. Pretty much as you’d expect, it shows exponential growth indicating a process of imitation. Note that the curve rises a bit above trend on Friday, but on the other hand it’s not entirely a Friday thing since you do see growth on Wednesday and Thursday too. I’m gonna split the difference and say it’s about half garden variety imitation and half the fact that Friday is the Islamic Sabbath.

Let’s hope the curve starts bumping up against the asymptote soon and goes from exponential to s-curve. On a more pessimistic note, even after this particular issue burns out, the tactic itself of drumming up outrage against an obscure blasphemy will be imitated at some point in the future by some political entrepreneur, just as this itself was almost certainly inspired by earlier similar efforts by other policy entrepreneurs. That is, there is a logic of imitation at both micro and macro, for protests within each scandal and for scandals imitating each other.

A caveat, I did my best to get the dates right but it often isn’t clear, even in the original news story linked in the KML file. Also, thanks to Neal Caren and Matt Frost for pointing to and showing me how to download the file.

Entry filed under: Uncategorized. Tags: .

After Surveys Climbing the Charts, ch 4

1 Comment

  • 1. Graham Peterson  |  December 10, 2012 at 10:24 am

    Is it generally assumed that cascades start with a valence-heavy node, or some other kind of purposive first mover? Clearly interest groups are doing the tipping in the application above.

    But generally speaking, couldn’t we get a cascade just from bits of sand gathering sufficiently (and stochastically) in some corner of a network, such that they hit a necessary threshold and tip a cascade?

    The question then would be: how to test the purposive first-mover hypothesis against the stochastic first-mover hypothesis.


The Culture Geeks


%d bloggers like this: