Alternate title: Hey look, I get to show you what I learned in that one statistics class I took in college!
Meet Swivel, the exegesis of bad science in a user-friendly interactive format. This site gives you several hundred datasets to play with, allowing you to graph one dataset against another. For example:
In the last 30 years or so wine consumption and violent crime in the US have been moving in opposite directions. Let's all get a glass and get less violent.
Click on the link above to look at the chart. The site even calculates correlation values for you to make the non-science seem even more real.
So what's the problem with using data to draw conclusions like this? Swivel doesn't tell you how the change in one indicator causes the change in another, thus failing to answer the question of what causes what. If two lines move in tandem, is the first causing the second, the second causing the first, or is it just a coincidence. Take my age for example. Just like clockwork (and as a result of clockwork) it goes up every year. Every year, the cost of a megabyte of storage space goes down. Almost every year, world GDP goes up. Every year, people pay more for baseball tickets. Does my age cause economic growth? Would my age stall if memory prices remained steady? Of course not, but Swivel wouldn't tell you that.
You raise some great points about human nature. We really like to 'draw conclusions' and assign causality. The cool thing about doing it on swivel is that it is fun. But for the discerning folks who do understand statistics and the idea of 'Correlation without causation', graphs and data like that can also serve as a great point of inspiration.
I personally never would have thought to investigate whether there is any correlation OR causality between wine consumption and violent crime. But the fact that the correlation is there can spark me to find out that perhaps there could be causality there.
I suppose - perhaps higher income means more wine and lower crime.