In recent days, we've heard lots of editorials masquerading as news. They tell us that the economic rebound, the successful rescue of the captain of the Maersk Alabama, and housing starts are due to the president's economic plan, which hasn't even hit the ground yet. We hear from DHS that subversive militias are recruiting former members of the military in an attempt to overthrow the government. We hear that the world hates America because we are greedy.
What we don't hear should shock you. What we don't hear is that the pirates are attacking convoys of food and earning $150 million per year in ransom money. What we don't hear is that the Commander of the USS Bainbridge gave a split second decision command to fire that saved the captain of the Maersk Alabama. What we don't hear is that compared to last quarter, housing sales are actually down, but it seems up because fewer homes are going on the market so the percentage is down (my realtor gave me the stats herself on this).
Last night in class, we discussed one aspect of bad science- the ipso facto logical fallacy. They assume that because certain things are concurrent that one causes the other. You can't say that. At best, evidence suggests a link, but on further experimentation we can easily dispell this. One example in class deals with a woman whose flashlight doesn't work. She changes the batteries and it works. What if it hadn't? What other things could it be? We ignore so many things because we do not think of them. My students came up with some good ones: corroded connections, burnt out bulb, batteries inserted incorrectly, etc. Just because new batteries don't fix it doesn't mean they're bad either. Sometimes it's a matter of compounding variables.
Another powerful example, since I live in Nevada, is the roulette wheel. Many novice gamblers assume that because black has come up ten times in a row that it's red's "turn" to get a shot and they bet on red. Truth is that on the next spin, there are the exact same chances of black turning up again, and they don't have abetter chance of being right about red. Unless you remove a number when the ball falls on it, it doesn't change the chances of subsequent draws. It is mathematically possible on an unaltered roulette wheel for a ball to land on 17 black EVERY SINGLE TIME, however unlikely that might be.
See one other problem lies in differences between circumstances and operators. Every person performs experiments differently. A coworker called me yesterday to ask me how I innoculate a specific culture because it only works 50% of the time for her and it has so far always worked for me (most things don't work that frequently). Variations and variegations influence outcomes. We cannot predict the future because the circumstances are NEVER the same in subsequent trials. Things unseen and unknown change all the time, so no matter how well we try nothing is ever an exact replicate. Not even identical twins share everything.
I spent probably 75% of my time in laboratory troubleshooting to pin down unexpected outcomes and link them to unaccounted variables. About half of the rest of the time, I had to explain differences between trials to justify omission of results. Some scientists get excited apoplexy when they "discover" something new, when more often than not they are ghosts and not a result of our manipulations. Just because two things coincide does not mean they are linked. Just because they are linked doesn't mean one is causative. Sometimes things just happen and there is no real good reason.
No comments:
Post a Comment