# Econstudentlog

## A few notes on (meta)rationality

So, let’s say you think policy X is optimal and policy Y is not. Or perhaps religion X is true and religion Y is not. Or you know something about subject X and you think you’re right, even though other people disagree. Now, if you’re like most people, you haven’t taken a closer look at the data.

Not necessarily, mind you, the policy data or the data supporting or questioning the religious ideas. Most people use some form of this type of data in their arguments, perhaps not as much because they find the data convincing but rather because they think they need to justify their beliefs somehow, and if you say that ‘policy X will result in more poor people’, or some kind of stuff like that, odds are that added information makes your position look more convincing to the opponent than if you chose not to say it. But the ‘unemployment will go up 2,4 % if policy Y is implemented’ is not the kind of data I was thinking about here. I was thinking about the data on who thinks what. Background variables. Do people who think X have stuff in common which might explain why they think the way they do? It’s an important part of understanding the subject – if your age or gender affects your opinion on the subject matter, disregarding those factors when explaining why you think the way you do leads to a potentially huge omitted variables bias. In short, it can cause you to deceive yourself about which factors have been important in the formation and development of your views. You think that you think X because of A and B (‘unemployment will go up 2,4 %’); but really it’s more a mixture of A, B, C and D.

People make arguments constructed like this: I think/like/prefer X because Y, where Y is some variable that pertains somewhat to the validity of the arguments under evaluation. Like, say, unemployment. Maybe I think the other guy’s argument is faulty or incomplete. Perhaps A (‘taxes’) is more important to me than B (‘environmental safety measure Q’). On net, the amount of supporting arguments in favor of X is higher than the amount of arguments in favor of Y. Things like that.

Here are some other things you might say in an argument – I don’t think most people bring up stuff like this very often, and when they do it’s mostly the characteristics of the opponent in the argument that gets the attention. To bring up this kind of stuff in an argument can go from being considered irrelevant to the matter in question to being considered an unjustifiable attempt to smear the opponent. The funny thing is that variables and related inferences like the ones below sometimes have extremely high explanatory power when you want to estimate what individual A thinks about subject X. We know this stuff matters a lot, but people really like to pretend it doesn’t and it’s often considered cynical or perhaps downright rude to bring it up in conversation. Here are some of them. Of course no one of these will have 100 percent explanatory power either, so I urge you not to reject arguments like these out of hand because they only explain part of the variation in the data – think of them as variables you might decide to estimate in an econometric model while trying to explain, say, the distribution of the opinion variable Z: