An example of correlation does not equal causation would be:
If Happiness catches, so must sadness, anger, and every other emotion. I Always feel pissy when people bitch and complain about stupid things...
In this single example, the correlation made, is that when other people are angry, I get angry. This is not actually the direct cause of my anger though. The reason for my anger is actually the incessant nature of the people complaining, and how silly the things they talk about actually are. Their anger doesn't actually bother me, it's the WAY they argue, and WHAT they argue about. Thus my correlation, albeit related to the cause, is not the cause itself.
Having said that... If someone tracks my behaviors, and statistically, when people get angry at something and start bitching, I actually start getting angry myself and bitch back, every time for different reasons each time, then they have proven their point with me.
If they can statistically prove this for hundreds of people, then they have a good theory.
If they can prove that regardless of what makes people angry initially, angry people make other people angry regardless... they have identified a phenomenon.
If they can prove that anger from one person, can cause someone else's anger inderectly through the people they come in contact with, then they have a significant phenomenon. For example, If they can prove that the anger caused, lets say by people bitching about stupid things on Slashdot, affects people who don't read Slashdot, a significant phenomenon has been tracked.
And if their statistics and theories can back up that significant phenomenon, then you have a significantly reliable theory. You will never be able to prove this theory 100%, but you will have something that is almost as good as 100% proof. A reliable theory (such as the theory of relativity) is useful until someone can prove it wrong... and even then, it gives us a damn good explanation so we can go on to move on to more important things instead of STUPID SHIT like this.
The original article did all these things. Except the slashdot-specific thing of course...
Now, if you'll excuse me, I'm going to stop talking now, because apparently this study just proved that it's very likely that I'm just pissing off all your friends, neighbors and children. And I don't think they did anything wrong in this argument...