Saturday, June 20, 2009

The Margin of Error

In the quest for scientific truth, there is little mention of the fact that during the course of controlled experiment or the gathering of data that the failure of the soda pop machine to dispense the right change to the research scientists altered the outcome of the research to a statistically significant degree. That explains why the space capsule landed in another hemisphere, upside down. Not only did the soda pop machine have such an effect, but so did the bacterially-active bologna and mayonnaise sandwich prepared by the kitchen staff. And the brittle cataracts in the researcher's eyes. And the phone call in the afternoon from the cousin in jail asking for bail money. This is the Wonderful World of Science. Ignore the convict behind the curtain, you are looking at Great Truths.
I think a lot about that phone call from the incarcerated relative, begging for lunch money and a pie with metal-file filling, as I walk about the boreal forest in northern Minnesota in search of things that in all probability do not exist but cannot be excluded without an all-knowing frame of reference, which frame of reference can be approximated through statistically significant sample sizes, generally numbering well less than infinity, which, doggonnit, invariably fail to consider that one sample that contains that which you, in the end, had assumed not to exist. Maybe if we maintain the sample size but enlarge the number of identical studies we can exclude the possibility. Or maybe if we have a massive amount of people say the same thing we can make it come true. Why, the sheer force of my personality might do it. But I am on a tangent and I am struggling to stay on point, as are the rest of my colleagues.
It was a lazy day about the woods and I thought it fine to engage in some informal postulations, perhaps stumbling upon some Higher Truth along the way. Make haste, I said, for a moment you can imagine that the cause of humanity rests on your shoulders. I observed: 1) I am a carbohydrate burning vessel. 2) I exude carbon dioxide 3) I attract carbon dioxide seeking organisms. Perhaps, I surmised, I can determine my "Carbon Footprint" by measuring exactly how many of such organisms were attracted to my person. Determining the total surface area I occupied at that moment, I arrived at 1.94 square meters, a figure, I discovered, that was identical to the amount of surface area occupied by the eleven heavily carbonated feral cats that lived behind our house in South Dakota in 1991 or the eight bottles of Fonseca Vintage Port 1970 in the wine rack. The wine caught the attention of a group of my colleagues, and all at once they set upon a rigorous regression analysis. For two hours, they carried out 16 repeated measurements involving eight independent variables. They found that after each measurement, the dependent variable approached zero. They were on to something. The unknown parameter appeared to be within grasp. They thought it might have something to do with the pizza delivery boy appearing seven times, clearly a random variable, but the exact function wasn't known. But they abandoned the study at the last measure; for all the while they had been carrying on they had found that their own surface area had expanded to a shocking degree, well beyond what anyone would have predicted. The phone rang and it was my cousin again.
Anyhow, a census of the black fly population that taxied about my surface area revealed 6,432 flies. This is some 3000 flies per square meter. Now to be useful, one might say that I need to measure my carbon output and I need to compare this to other carbon producers. But I have already determined that these are 6,432 random variables at any given moment in an environment where I, of 2 square meters, am in search of a species which depends upon my observation for its survival. This search is through third-growth recovering forest, stripped of old growth characteristics, choked with aspen clones and hazelnut thickets and mountain maple and balsam fir deadfall, all observed through the matrix of mosquito netting and the haze of fogged prescription glasses, beaded with rain and perspiration, while an electrolyte-depleted circulatory system produces leg cramps and heat exhaustion and iron streaked rocks send the compass spinning like a roulette wheel and the black flies sound like rain as they bombard my clothing and this is supposed to determine the nonexistence of an object in a 50-acre patch of forest.
The formula should read something like:

Where Y is the rare plant frequency and X represents independent variables including survey intensity, stand potential, phenological stage, light intensity, deadfall proliferation, drought index, spruce budworm kill factor, seral stage, deer density, slope, soil moisture, wind speed, surveyor education, surveyor experience, surveyor organizational skill, surveyor lactate levels, electrolyte imbalance, neurotransmittor depletion, excess body temperature, eyeglass opacity, cornea deterioration, cognitive disassembly, caffeine-induced confidence, memory loss, methamphetamine lab density, mosquito netting shear strength, boot porosity, pencil loss, blister quotient, degree of disorientation, fungal growth rate, bone fracture, anxiety level, battery failure, life insurance dollar amount, declining profit margins, and bitter regret. Each E represents one of 6,432 black flies.
I think I am on to something. Here we begin to figure out the value of B. The dependent variable Y is inversely proportional to the value of the independent variable X and the error factor E. As X and E increase, Y decreases. So let me get this straight: This is to say, as the probability of the existence of a rare species approaches zero, so do I.

From northern Minnesota:
Dryopteris fragrans, or Fragrant fern. Rare in WI, but common in MN. Very pleasant smell when crushed. Found on sheltered, vertical cliff face, typical habitat.