Well, with little planned in the way of textual posts (although I have plenty planned for other types of posts), now is as good a time as any to start posting the second half of my political analysis notes. I already posted the notes from the first half of the semester, which you can find grouped together here. Those posts are all about power. These coming posts will be from the second half of the semester, and will focus on aggregation. So, let’s begin.

The Latin word “grex” means “flock,” and “ad” (which becomes “agg” in aggregation) means “to,” so “aggregation” is assembling a flock. It’s clear how power pertains to politics—but how does aggregation relate to it?

Well, let’s start with another question: Why war?

Maybe it’s a spiritual problem, as the Dalai Lama would assert—a problem of misunderstanding and hatred.

Other people believe it is a diversionary tactic—leaders need support of the selectorate (the critical sectors of the voting society.) So when there are domestic problems, the leader will create a foreign policy crisis to distract the electorate and unify them against the common enemy.

Then there’s the strategic theory. The strategic theory says you need two states for a war, so it can be modeled as a game. And as we’ve seen, if people don’t trust each other, they’ll end up with suboptimal results—war. Another strategic theory is that war comes about when there is incomplete information—both sides are unsure if they can win an armed conflict, so they have to duke it out to find out, rather than relying on the validity of each other’s threats.

Finally, there’s systemic theory. Systemic theories don’t look at individual states or dyads (pairings), they look at the system. There are three types of systems—the one, the few, and the many.

The one, or unipolar, is a system with a single strong state, like the Aztec Empire in Mesoamerica, or the US since 1989. Also Childhood’s End.

The few, or bipolar, is a system with two powerful states, like Athens vs. Sparta, or the Cold War.

The many, or multipolar, is a system with more than two powerful states, usually a lot more than two. Examples are the concert of Europe, the Warring States Period of China, or A Song of Ice and Fire.

And if you want to determine if a war will happen, or why a war will happen, you have to determine what kind of system is present.

Hobbes said unipolar systems are best, because everyone will be subject to one power which will ensure that there’s no infighting. We’ll talk more about this later.

Neorealism disagrees with this, arguing that bipolar systems are better, because the two powers will be in competition, and both will work harder to ensure that what they control is peaceful.

And Wilsonian idealism champions multipolar systems, and a peaceful league of nations.

So, why all these theories? Which one is right?

That’s tough, because they’re all explaining war at different levels. The spiritual explanation looks at individuals, where diversionary theory looks at a single nation, strategic theory focuses on dyads, and systemic theory looks at the whole big mess.

So how do you link these levels together? How do you aggregate them? This, to some extent, is our question.

The Ecological Fallacy

The most basic form of aggregation is just to put everything together. Put together a bunch of individuals, you get a nation. Put together a couple nations, you get a dyad. Put together a bunch of dyads, you have a system. That’s what Professor Dion calls the “rock pile” method. If you put together a bunch of rocks, you get a rock pile. So if you have put together a bunch of dumb people, you’ll get a dumb group. If a majority of republicans are elected into congress, you’ll get a congress that votes republican.

But this isn’t actually the case. When you put a bunch of individuals together, things get weird.

To explore this, we’ll look at another question, similar to “why war?”: Who voted for the Nazis?

The problem is, they used secret ballots, and this information isn’t readily available. We could look at precincts and see how they voted, and what their demographics were. From that we can find a rough correlation between certain groups and voting patterns. This is perfect, right—or as close to it?

Not necessarily, because it’s an ecological fallacy. That name is a bit misleading, so I’ll explain. It comes from sociologists applying biological sciences to individuals—moving from individuals to whole groups, just like biologists moved from species to ecological systems to better understand the individuals. So sociologists applied this ecological approach, using the same sort of precinct analysis above, but this method was proven ineffective and inaccurate.

To summarize, the ecological fallacy involves using information from one level of analysis to make inferences about another level of analysis.

An example:

A Berkley graduate admission study in the 70s found that 46% of men were admitted, while only 36% of women were admitted. An ecologically fallacious argument would be that the school is biased against women. But by looking at each department, they found that some departments actually heavily favored women, while others just slightly favored men. So what happened?

Men applied to departments with high acceptance rates for everyone, while women applied more often to departments with lower acceptance rates. So it wasn’t that UC Berkley was discriminating against women—it was a problem of analysis on one level versus another.

Next post will look at a more accurate use of aggregation—the Condorcet Jury Theorem.