[quote]AssOnGrass wrote:
dhickey wrote:
Sloth wrote:
Thanks to Newt!
Probably thanks to the telecom bubble.
Stop throwing out inconveniences to his argument.
Why he hasn’t acknowledged this is beyond me.[/quote]
You guys gave me 4 pages of posts to address. I am extremely busy and I’m sorry if some of you feel like you have been ignored. I haven’t ignored you.
The fact of the matter is that your argument is fallacious.
Here’s why:
Let’s assume that unemployment and presidency have nothing to do with each other. If that is true, than unemployment will go up or down regardless of who is in office. If they are truly independent, a given president has a 50% chance that unemployment will be higher than when his term began, and a 50% chance that it will have fallen.
Let’s take this a little farther. If unemployment and presidency are independent it is still possible for unemployment to decrease for a democrat and increase for a republican. There 50% chance that one presidential period will follow this rule. There is a 25% chance that two consecutive presidential periods will follow this rule. For three periods, the odds are 12.5%.
If unemployment and presidency are truly independent, the odds that unemployment will decrease under all blue periods and increase under all red periods since 1900 – 11 periods total – are %00.048. A fraction of a percent.
So yes, it is possible that you guys are right and I am wrong. But it is extremely unlikely.
The much more likely scenario (99.962%) is that unemployment and presidency are somehow dependent on each other. In this case, there are three possible causal links that could be at work, but the interesting thing is that these do not need to be understood in order to make use of the predictive power yielded by correlative evidence.
The specific circumstances surrounding each period of rising or falling unemployment do not need to be discussed. It will always be possible for you to back-rationalize and pick and choose this and that factor that you want to claim caused the rise or the fall. (See complexity theory and information theory as applied to beyesian modeling if you wish to understand why). There is no inherent validity in this method unless you can use your ideas to make predictions.
If you are descriptive and not predictive, you are not scientific, but you will fit right at home amongst economists and other followers of dubious fields that are ridiculed by the hard sciences. Unfortunately, economic theory is too often not held to the same light of rigor found in other fields and many economist do not grasp the basic underpinnings of the scientific method. The relationship between correlation and causation is one of the most basic tenets of the scientific method.
If you do not understand that correlation holds predictive power even in the absence of any assertion of causality, you would not make it far in any field that professes itself to be a science except economics. I question whether there is anyone else in this thread who understands this.