Monday, February 29, 2016

Sharpshooter: the proud fallacy

Suppose you’re a cowboy. Now suppose you’re shooting a blank canvas. You’ve shot the canvas about fifty times so random clusters are starting to show up. Your friend calls you and tells you he’s on his way to see how you’re doing. You do the only thing a logical, proud cowboy would do-paint a bullseye around your best cluster. 

Congratulations! You’ve just committed the Texas Sharpshooter fallacy. Any time that you use the same data to both construct and test a hypothesis, you’ve used this fallacy. This fallacy blurs the line between evidence and data that might have just been random, and making it appear to form a causal relationship. 
Now I have to admit that after researching and coming to understand this fallacy, it seems that I have used it in the past. I was writing a paper over the significance of violent pornography and its affect on children. I looked up tons of different research sites, chose a couple, and then formed my hypothesis. This backwards thought process not only skewed my data, but made my paper seem a lot stronger than it would have been if I’d solidified my hypothesis and then done research. By choosing only the sites that strongly supported my argument, I misled the reader. 
I believe that there are different ways that this fallacy can come into play. 1) If you pick and choose data to support your argument when there is a myriad of conflicting data, you’re creating a bullseye and therefore using this fallacy. 2) If you tweak your hypothesis after you’ve done research and realized you weren’t quite correct, you’re moving your bullseye and therefore using this fallacy. 3) If your hypothesis is constructed using the same data that you test it with, you’re painting a bullseye after the fact and therefore using this fallacy. 

This fallacy is one that is all too common and easy to utilize. The goal of writers is to create as strong an argument as possible and by understanding this fallacy I believe that my arguments could have been stronger in the past. The fact of the matter is, if the data doesn’t correlate in a significant way, maybe your hypothesis is wrong and you should start from scratch. I wish I’d had this class in the past to teach me the simple rule of setting aside pride to come to a concrete answer.


No comments:

Post a Comment