That's very interesting and the mathematical model part seems great. But shouldn't you have some conception of what constitutes persuasion or political support before including it in the model? After all, if it's not something we have an independent grip on then why not just phrase the model in terms of persuasion/support expenditures. And I worry about thinking one has a grip on such notions without a specification of what sort of thing one means (may not always be the same in every use) and whether the simplifying assumptions one made are valid on that interpretation.
I mean it's really easy to think one has an idea of what it means for there to be more or less propoganda but those empirical results you mention immediately raise questions. If media competition in Russia results in more media outlets -- all of whom engage in some level of government propoganda -- does that count as more or less propoganda?
I'm a huge believer in the value of these kinds of simple models but constructing the model and deciding how to understand terms in the model in terms of the actual world aren't independent. I see perfectly good models confusing people because the way the model treats a term (eg persuasion or support) requires it be measured in some counterintuitive fashion.
For instance, our intuitive conception of political support doesn't satisfy the assumption that it's expensive -- failing to offer it during a war is often expensive. But if support is something more like coming out to vote it might satisfy that assumption but you better be aware of that difference because now support has actually decreased when something is do uncontroversial few people bother to show up to vote.
Also, I think the assumptions you made to reach conclusion 5 are somewhat unclear hence some apparent conflict with observation. Competition in markets often improve the products -- if the federal government provided the only sodas they'd likely suck -- and often a dollar spent on persuasion is more effective if the product is better (easier to make me thirsty for coke than beat juice).
That's very interesting and the mathematical model part seems great. But shouldn't you have some conception of what constitutes persuasion or political support before including it in the model? After all, if it's not something we have an independent grip on then why not just phrase the model in terms of persuasion/support expenditures. And I worry about thinking one has a grip on such notions without a specification of what sort of thing one means (may not always be the same in every use) and whether the simplifying assumptions one made are valid on that interpretation.
I mean it's really easy to think one has an idea of what it means for there to be more or less propoganda but those empirical results you mention immediately raise questions. If media competition in Russia results in more media outlets -- all of whom engage in some level of government propoganda -- does that count as more or less propoganda?
I'm a huge believer in the value of these kinds of simple models but constructing the model and deciding how to understand terms in the model in terms of the actual world aren't independent. I see perfectly good models confusing people because the way the model treats a term (eg persuasion or support) requires it be measured in some counterintuitive fashion.
For instance, our intuitive conception of political support doesn't satisfy the assumption that it's expensive -- failing to offer it during a war is often expensive. But if support is something more like coming out to vote it might satisfy that assumption but you better be aware of that difference because now support has actually decreased when something is do uncontroversial few people bother to show up to vote.
Also, I think the assumptions you made to reach conclusion 5 are somewhat unclear hence some apparent conflict with observation. Competition in markets often improve the products -- if the federal government provided the only sodas they'd likely suck -- and often a dollar spent on persuasion is more effective if the product is better (easier to make me thirsty for coke than beat juice).