Idle Thoughts

7.26.2010

Numbers

I haven't delved deeply into the study of statistics, but I know enough and I feel like I know more than most people.

There are three kinds of lies: lies, damned lies and statistics.

That quote is the best place to start. A statistics reduce the world to a relational-quantitative perspective. If I tell you that I have 3 apples, I'm reporting a statistic. If I tell you that I have 1% of the apples from my neighbor's tree, I've reported another statistic and providing you with the information needed to infer that that the yield of my neighbor's tree was 300 apples. If I tell you, 1% of the apples from my neighbor's tree were reported stolen, you might be tempted to guess that I stole them; the numbers match up, 3 were stolen and I have 3 in my possession. That is a coincidence, yielding circumstantial evidence.
That was a rudimentary and facetious example, but my point with it is, statistics don't lie. They can, however, be used to lie. Although, the lie itself comes from shared ignorance, as easily as malice.

Percentages are a particular and common source of deception or misunderstanding. The power of percentages is that they make large values comprehensible, while highlighting relationships. To see the danger, requires understanding that any number is, itself, an abstraction of reality; a percentage is a second degree of abstraction. An immediately apparent danger stemming from this, is that (for example) 33% may mean 1 out of a sample of 3 or 330,000 out of a sample of 1,000,000.

Then there are the common sorts of mathematical errors associated with percents. For example, there's a piece of information that makes the rounds every so often about the relative salaries of men and women, which asserts that women make 75₵ for every 1$ a man makes. Seemingly nice, hard numbers. Usually this is part of a rant about gender inequity and the writer goes on to assert that men make 25% more, which is wrong. Women make 75% of what men make, which is 25% less than what men make, but men make 33% more than women. This because, the 25₵ difference is 1/3 of the 75₵ women make while it is also ¼ of what men make. Simple mathematics, but it shows how poor we are, in general, at correctly seeing simple mathematical relationships.

It's common to see increases and decreases miscalculated in that fashion, as well as errors with multiplication, division and even choosing the right operation. Which is the other reason that it's necessary to be skeptical whenever you're presented with percentages sans hard numbers.

From here, there's the problem of probability and it's necessary to acknowledge evolution. An essential capability of animal life is the ability to assess probability. It's fundamental, to animal survival, to recognize correlation in the location of food source, in the hunting behavior of predators, in the evasive behavior of prey. It's innate within us to recognize pattern, there's even a certain thrill or joy in it (which I suspect is an underlying motive for art and science).

Where the human mind deviates from a common animal, is with the ability to step away from intuitive pattern recognition, in search of reason for the pattern. We are able to ask “why is this so?” Which is a powerful question, enabling us to distinguish cause from effect and at times to manipulate the world to our benefit.

Conflicts do arise between our intuitive sense of the world and our conscious examination of it, this is typically a limitation of perspective. As example, we walk about a roughly spherical body of rock, which due to scale, seems to be a flat body. This rock orbits a star, which seems to rise on one side of our flat rock and set on the other, while in fact, the Earth is itself, rotating about it's axis. While the Sun is not visible, a moon and other fainter stars seem to orbit our flat rock. We seem to be the center of creation.

The flat and Earth centric hypotheses are seemingly reasonable, through the limited perspective of our ancestors. Eventually, evidences accumulated, proving first a round Earth theory and later a heliocentric solar system theory, which has all combined into our modern understanding of the universe.

Maybe I digressed too far but the point is, the mind is limited by perspective and experience. Meanwhile, we do have a built in sense of probability, an intuitive sense of a concrete reality but we need to develop a mindful awareness of the limitations of our perspective. Science does this and takes it all a step further, looking for ways around the limitations of humanity.

Knowing that, our bias affects our ability to accurately asses probability, we need to be cautious not to conflate the personally common or rare experiences with the real likelihood of an event.

Role Playing Games offer a good example for this: when the player kills an NPC (Non-Player Character) in these games, items are dropped. The real odds of a particular item dropping may be 10%, meaning each kill brings with it a 10% chance of attaining the item, regardless of how many times it's been killed, or who has done the killing.

It's reasonable to assume the item will drop within the first 10 kills, there is a reasonable probability it will drop on the first kill, it's even possible (tho not likely) that the item will not drop in 1000 kills. Still, there are players who go in knowing the odds and call BS if they don't have the drop by the 10th kill, while others go in blind but get the drop on the second kill and then deduce there was a 50% drop rate. Both types post on websites about the RPG with their personal testament, misleading or confusing other players.

Another error is to assume that things which seem astonishing are improbable or impossible. A common argument for creationism asks, “what are the odds that our moon would just happen to be the right size and at the right distance to perfectly eclipse the sun?” Ignoring the truth, which is the moon is not perfectly sized, shaped or positioned, the reality is the common person lacks the knowledge to assess those odds.

The modern, naturalistic hypothesis, supposes that 2 planets formed in near or overlapping orbits and at some point (relatively) early in their existence, the smaller one (roughly the size of Mars) collided with the larger one, ejecting material that would coalesce as our moon.

What are the odds of all that? The honest answer, and the answer that would come from a person with a layman's understanding of statistics and the universe is, “I don't know, but the universe is so vast, that even the exceedingly improbable stands a chance of happening, at least once.”

There are other honest answers to that question. A natural universe, is to some degree and on some levels, deterministic, in that matter interacts with itself and space in particular and specific ways. I think it's safe to infer, our solar system began in such a way that favored and would lead to the current planetary alignment.

1 comment:

  1. Just managed to read this one. Double thumbs up. Its a very insightful piece. I'm not sure what role statistics plays in my life, but I have noticed I think about it far more than many others in my life. In any given day I consider the statistical probability of at least 5 events in my life.

    ReplyDelete