I guess this ends the gun debate once and for allFollow

Gathering more information doesn't make it science anymore? That's a new one.

In my example, he traveled to the "new world". Yes. I'm aware of the historical confusion regarding where Columbus actually thought he went and where he ended up. Um... The point was that he didn't fall off the edge of the world getting there.

Good example of liberal obfuscation though!

Read what I wrote.

Gathering more information purely because the information you have doesn't match the conclusions you want to arrive at isn't good science. Doubly so when the methods you use to gather more information are designed to feed that starting bias.

Gathering more information, regardless of the motivation, is good science.

I really don't see what you're afraid of. If you truly believe this finding to be accurate then you should welcome more data to prove your conclusion.

Okay, I lied. I know exactly what you're afraid of.

Decreased gun control regulations in the US consistently results in decreased overall crime rates. The other variables are largely meaningless because this effect tends to occur regardless of which of those other variables are present at the time. The gun control folks keep pointing at the variables in this case, and that, and the others, but like to gloss over the fact that it's not the same set of variables in each case.


To put this in math terms, it's like the following sets:

A+B+C+D=1

2A+2B+.5C+.5D=2

2A+.5B+2C+.5D=2

2A+5B+.1C+2D=2

2A+1E+2F+1G=2

2A+3E+.5B+4F=2


etc...


What we see is that when we double A, we double the result. This occurs regardless of the changes in other variables. A good scientist would conclude that the weight of A in these sets is vastly greater than the weight of every other variable we're including and conclude that A is the most direct cause of the result.


A bad scientist would insist on focusing on the other variables, and attempt to discount the first guys findings because he didn't account for them all.


Some people understand this principle of significance. Most don't. That would be you guys btw...

Edited, Mar 30th 2010 1:08pm by gbaji

But what we have, so far, is two possibly related variables and no methodology.

Gun ownership increases. Violent crime declines. Are they related? We don't know. We don't have enough information, and therefore to a person with an agenda to push, information is the enemy.

Ice cream sales increase. Burglary rates increase.

Does ice cream somehow encourage burglary?

So simplistic answers ignoring possibly statistically significant variables are best?

Honestly gbaji

We have a host of methodological examples though. That's the point.



Added the relevant bit, just in case.

Here's the thing. Crime rates decline in virtually every single case, regardless of other factors. Urban areas? Decrease. Rural areas? Decrease. Different ethnic makeup? Decrease. Southern states? Decrease. Northern states? Decrease. Tornado areas? Decrease. Hurricane areas? Decrease. Bible belt? Decrease. Non bible belt? Decrease. High drug use? Decrease. Low drug use? Decrease.


There's a point in all of this at which constantly pointing at all of those other factors and insisting that we can't make any conclusions without considering them all becomes silly. We can exclude them to some degree because they don't seem to have much impact on the effect we're looking at. The relevant data is overwhelmingly in favor of loser gun control regulation. That's exactly why those who favor stricter gun control resort to tossing in irrelevant data in the hopes of confusing the issue and allowing them to insert rhetoric and emotional appeal where logic and facts don't work.


Let me also point out that even in a "we don't know" situation, shouldn't we *not* infringe a constitutional right? Shouldn't we only do that in a "we know absolutely" situation? I should hope so...




If you had just one example? And with no rational explanation for a connection? No. But when the correlation works out that way consistently and in a wide assortment of cases and there is a very rational explanation for why said causation would occur? The answer becomes "yes".

It's the "possibly statistically significant" part that the gun control argument has problems with though. You can't just keep lumping in data points which "might" be significant. At some point, you have to show that they are. But the gun control argument doesn't do this. It just leave hanging the possibility that the data they keep tossing on the heap might be significant.

That's not good science. Good science requires that one weed out the insignificant in order to determine the significant. The opposite is done when you don't want to know the real answer, and hope to prevent others from figuring it out either.

In chemistry, if you're trying to identify which parts of a solution are causing a reaction you're seeing, you remove them and keep testing, right? The guy constantly saying things like "Well. What if we add benzine to the mix? Or Chlorine! Or... Whatever's in this bottle! What would happen to the reaction then?!" isn't really helping, is he? While you certainly might gain extra data doing that, you don't learn which of the chemical agents you were looking at is causing the reaction. You reduce the complexity, you don't add to it.


That's how science is done. You take things away to their simplest parts. Doing it the other way around makes it easier to speculate on what might be, but it does not help us determine what is.

So what you're saying is that you need more data?

But not so much data that it changes the initial results. Then it's liberal obfuscation data.

And honestly, that would be legit. But, and this is important, it's legit no matter what your bias.

I know, and nothing has changed. I still gagged at what he wrote.