Trust the Ferret, the Ferret is Wise

So now I'm studying engineering, which means I'm studying more advanced math, like linear algebra and differential equations. Which are hard, or at least Diff EQ is. But then, math is a game for the young;.

But that's not the thing I'm noticing.  I'm noticing things where the different math I've learned mets and intersects, and works together.  And because this is engineering, how it works to predict and calculate the real world. Which sometimes seems really damn weird.& I can sit here with a pencil and paper and predict all sorts of physical things to decent degrees of accuracy. Why should it be that we can figure things like acceleration and stress out by putting numbers in columns and then messing with them? Take numbers and just mess with them and the answers fall out.

Now, the obvious answer to this is that math works to describe our physical universe, because it was created by people living in the physical universe, so naturally enough the rules we develop for mathematics are going to match up to the way the universe works, since we're using it to describe the universe. At least if the universe is broadly comprehensible and follows rules of cause and effect and repeatability. If it doesn't, then all our effort at trying to find rules is just us finding patterns that don't exist in chaos.

The problem with that is what about mathematics that doesn't describe things we can currently observe? Nth dimensional algebra, fractal spaces, and the most rarefied bits of math that don't seem like they have any relation to the universe? Do they represent things like strings and exotic matter bubbles? Or what?And if they represent those, how freaking weird is it we can figure out the rules before we find the things? And what about the N dimensional universes we can describe that may or may not be correct? Do they describe the non-spaces where other entities exist, which we would describe as squamous and rumose?

(Don't mind me. I've been reading Charlie Stross. But this kind of feeling does hit me every so often, it's just so WEIRD sometimes to do these complicated math to numbers and find out that yes, this reflects something that really happens)

Higher math really just doesn't feel fair. It feels like I just take the different numbers and then just mess around with them until they come up how I want. There's rules and things, but most of it just feels like I'm tossing stuff together however I want.

There's a good bit that's happened in the past couple days, but I can't really say anything you guys haven't already seen about Scotty, or about London, or other things. So I'll continue my rambling series of things where I lecture.

To start with, the topic of this one's a lie. Statistics aren't worthless, they're very valuable, if used right. But, just like most kinds of birth control, if they're used wrong, they're worse than useless. For way of example, an excerpt from a bit of radio I caught yesterday. There was a guy on, saying how 42% (I think, there was static) of Christian kids had watched MTV in the last (something), as opposed to like 59% of non-Christian kids. Now, he started to go into things about having half of your youth group stand up to show how many kids did it and so on, but static and boredom overwhelmed me, so I changed the channel. But there's a bunch of things that we can glean from his use of statistics. Which is, frankly, poor, not to mention deceptive.

The first thing is the underlying assumption ions. Sure, there's the one about watching MTV is anti-Christian, or not something "good Chrisitains" do. Now, I'm not Christian myself, but I grew up attending church, and I've read the Bible, and don't recall MTV being mentioned once. Jesus tended more to focus on things like the poor, and the moneychangers in the temples, and the Pharisees, and smiting fig trees. But that's not the only assumption. He's also assuming all Christians believe the same as he does, which is manifestly not true. There's plenty of Christians who have no problem with MTV. More than 70% of the US population is Christian (see here ), that's 150some million people. Heck, the survey results quote
"Christianity. Note that in the NSRI and ARIS studies, based on self-identification, Christianity includes: Catholic, Baptist, Protestant, Methodist/Wesleyan, Lutheran, Presbyterian, Pentecostal/Charismatic, Episcopalian/Anglican, Mormon/Latter-day Saints/LDS, Churches of Christ, Jehovah's Witness, Seventh-Day Adventist, Assemblies of God, Holiness/Holy, Congregational/United Church of Christ, Church of the Nazarine, Church of God, Eastern Orthodox, Evangelical, Mennonite, Christian Science, Church of the Brethren, Born Again, Nondenominational Christians, Disciples of Christ, Reformed/Dutch Reformed, Apostolic/New Apostolic, Quaker, Full Gospel, Christian Reform, Foursquare Gospel, Fundamentalist, Salvation Army, Independent Christian Church, Covenant Church, Jewish Christians, plus 240,000 adults classified as "other" (who did not fall into the preceding groups)." That's a lot of variety, and a lot that WON'T agree with his views, they range the full spectrum from liberal to conservative and back.

That's most of the US. Of the other ~30%, you've got a mix of Jewish, Muslim, Buddhist, Hindu, Agnostic, New Age, Scientolgist, Atheist, and Lazy. (At least, I'm not sure what to make of the percentage of people who say they're "Nonreligious/Secular" but aren't agnostic or atheist. Just don't care, or don't want to answer, would be my guess. Or are "spiritual, but not religious". Anyway.)

So there's another flaw in his statistics, he's comparing one group that's more than twice the size to another group, and both groups are very heterogenus. Plus with no mention of the methods or margin of error of the survey, you can't really treat his results as anything other than garbage. The group he's trying to poll is too big and too varied for his statistics to prove anything, because there's too many branches of Christianity to generalize about them that much. And the underlying assumptions he makes, both about culture and about the people that were surveyed, are wrong. Which undercuts his whole point.

Another point that should be mentioned is how he cited the Christian/Non-Christian numbers without any mention of the size, there's a lot of folks out there trying to make it seem like Christians are some kind of persecuted minority in the US, rather than making up >70% of the population.

This is all to illustrate not only his absurdity, but the most important part of dealing with surveys, you have to look at them in context, and look at the data and how it was gathered. You can't just trust some number thrown out there without any references. And, to be honest, even the most rigorous surveys tend to have decent margins of error, plus most are voluntary, so there's a lot of people who don't answer. Information from surveys and statistics can be valuable at the big picture, but when you narrow them down, they lose a lot of value. And when it comes to an individual, they're almost worthless. Which is sort of like quantum mechanics, except completely not.

So what makes good statistics? Large sample sizes, and repeated samplings. The more data points you have to draw on, the more likely you are to get somewhat accurate averages, and things won't be skewed by the extremes as much. But that doesn't work in politics, since the extremes are the most active and vocal, and have sway that outweighs their numbers. Or money, which does the same.

Technorati Tags: Mindscribbles, SCIENCE!, Math