From Private Attitude to Public Opinion: A Dynamic Theory of Social Impact Nowak et al 1990

Nowak paper, not the game theory/evolution Nowak with all the proofs. Another social modeling paper, this one from 1990. it proposes a very pixelated 2-D model for social influence and persuasion, as usual abstract enough to model disease or fads. Lots and lots of paramters.

i liked the argument for 'reductive simulation' that is a good synonym for agent based modeling. essentially, instead of coding 'public opinion' as a digit 1-100, you code individual opinions and measure public opinion from that. You can see it as a multiscale model. Of a 2 scale model. 1-D, 2-S.

"According to Petty and Caccioppo (1986), when people are heavily involved in an issue, arguments will be processed centrally, and persuasion will depend on their relecance and quality. On the other hand, when personal involvement is low, peope do not pay full attention to messages, and persuation is deterined by such peripheral cues as ...."

"processed centrally"! interessant!

i made a mistake, this was the paper that distinguished supportiveness and persuasiveness, not the 2005 paper about innovation.

"we adopted the theoretically neutral assumption that ..parameters have a uniform distribution"

is that theoretically neutral?

They also say that the lattice is theoretically neutral. My note was 'Ixnay!', but that understanding was still 8 years away.

much influence from Latane 1981

supported by/ predictive of empirical observation that segragated communities are more likely to generate extremist opinions (Greer Orleans 1964). (I'd object to the word extremist, which is normative, but the idea is right, less drift with heavy coupling. I'm glad that is a sociological result)

Cites Gleick at the end, making the case that complexity can come from simple systems.

More 'case of the modeler': "

Computer simulation can serve as the test of a theory.  Writing the simulation program is a test of the theory's completeness and lack of interncal contradiction.  The assumptions of the theory seem more valid when they produce phenomena that are known to occur in social reality....

Computer simulation may reveal emergent properties of a social system stemming from laws assumed to opeate on the individual level.  ...we nood not assume group-level processes to explain them.  For example, our simulations have shown that no special forces attracting people of cimilar attitudes to move colser together must be assumed to explain group coherence.  Likewise, no special preces of greater majority persuastion is required to explain greoup polarizaiton, nor any notion of greater minority incluence to account for the fact that polarication is inomplete.

"

This sentence: "The assumptions of the theory seem more valid when they produce phenomena that are known to occur in social reality." is dangerous. Your conclusions do not validate your assumptions. that is circular.