I want to start by mentioning some things that struck me from this chapter. Firstly, Meadows’s augmentation to the Ten Commandments, “Thou shalt not distort, delay, or withhold information.” No, I don’t mention this out of some desire to “fact check” one’s ability to alter the Ten commandments. In fact, I encourage cranking things to 11. The reason this is an important thought is that we live in the information age. A fact that I remembered today when I asked Siri where the phrase, “Sh*t, Shower, Shave,” originated. However, to my dismay, she responded with: “Samuel! Watch your language!”
Information is easy to spread and to access; likewise, information is incredibly easy to falsify. I learned this in my days as a mad scientist. Whether we were out in the field, or in the lab, every data entry was vetted by another member of the team. Always. Our methods were carefully choreographed and executed, and our eyes were shielded from the data until all the analyses could be run. Because no matter who you are, you will search for data that points in favor of proving your hypothesis.
This brings me to the second quote that I found especially interesting in this week’s reading: “we don’t talk about what we see; we see only what we can talk about.” Meadows’s quoted an excerpt from a journal article written by Fred Kofman. I wanted so badly to have my data support what I had been talking about, and hypothesizing. Alas, data doesn’t lie. I had been measuring the inner chamber volumes of fruit fly galls, the corresponding hosts’s femur lengths, and the number of eggs the females held in their abdomen (fecundity), for the better part of three weeks. If the flies were bigger, it wouldn’t be worth mentioning, however, the flies were so small that it required peering through a dissecting microscope for hours on end. The process was slow, and tedious. Worst of all, it required that I drink no coffee in the morning in an effort to reduce the amount my hands shook while taking measurements.
After my data was gathered, I began the preliminary analysis on Xcel before moving my data into R. The initial results were extremely positive, with one R^2 of .86. I was sure that my hypothesis that the length of the fly femur length was in direct correlation with the gall’s inner chamber volume. I immediately began working feverishly on, what could have been, a publishable paper. However, a few months later (after my first draft was near completion) my statistics professor asked for me to run the analysis in R. The request on its own was expected, however, she added the variable of gender to be run in my analysis.
Before I continue, I should mention that like nearly all animal species, these flies had a level of sexual dimorphism. Which in this case meant that the females were significantly larger than the males. So, after running the further analysis, I found that those large R^2 values were not caused by a correlation of femur length to gall chamber volume, but were instead due to the sexual dimorphic nature of the fly.
In English: I proved that female flies are in fact bigger than males. D’oh!
I, too, encourage cranking things to 11.
Posted by: Cody Janousek | 02/16/2017 at 11:36 AM