Violent Media subtitle: Learn Math!

[Phopojijo]

Short little rant that annoys me... people really don't understand Math all that well.

The worst part is those are the same people who pull statistics out of their ass.

*****

I'll give you an example:

"Most spree killers have played violent videogames (or various other media... I'll just write videogames from here-on-out to be short) at one point or another, so we must cut down on them!"

****What's wrong with this logic?****

Their data states that spree killers play violent videogames... that is generally speaking true. There are exceptions, but statistics doesn't require perfect correlations to show causality, and vice-versa -- a perfect correlation does not prove causality.

Sounds good?

****But let's break down their logic:****

-> Their independent variable is a selection of people, controlled to only include violent people.
-> Their dependent variable is the amount of videogames they play.

Note that an independent variable is designed to be controlled to see a change in a dependent variable... so you're allowed to control the independent variable to only include violent people.

Your rate is therefore dependant variable over independent variable -- rise over run... good ole Grade 9 math.

Your rate is videogames played per violent person.

If you wish to minimize that rate, you should minimize the access to videogames. (Minimize the rise to minimize the rate)

****What does minimizing that rate do?****

Minimizing that rate reduces the access to videogames to violent people by reducing the access to videogames to everyone.

****Why does this feel awkward?****

I'm minimizing a dependent variable... ... but a dependent variable cannot be minimized since it cannot be varied... uh... hmm...

****What really is wrong with the logic?****

The independent variable should be access to violent videogames, and the dependent variable should be the amount of violence recorded.

I don't give a damn what videogame a violent person played in his spare time... I care that he is violent

As it turns out... according to the Washingston Post... with the independent variable set to Internet Access or Violent Media Access -- Violent Sex Crimes and Violent Physical Crimes reduce -- respectively. (http://www.slate.com/id/2152487/)

...
...
...

It's simple math people... Dependent and Independent variables are taught in Junior High School (depending on jurisdiction)...

People just really really do not understand how math works... and that's a fundamental problem with society. If more people were analytical like this, a lot of political bull**** would simply disappear.

...
...
...

Discuss?

[Sel]

So, math causes violent killing sprees?!

[Needles]

So we're not using all the independent variables (we're stereotyping on one kind of person), and for those who are violent cut down on the depend variable (the game variety?).

If that's not right, then idgi and I'll go with what sel said.

[Phopojijo]

No...

Your data defined your independent variable as the type of person then defined the type of game as the dependent variable.

You cannot vary your dependent variable(s) since by definition they are the resultant of some process or function on your independent variable(s).

*****
In a much-overused analogy:

Data: You turn on the sprinkler, you notice the grass got wet.
Rationale: If you notice the grass is wet you MUST have turned the sprinkler on.
Logic fails because: It could have rained, you could have let the dog out, it could be early morning dew...
*****

In other words:

Independent variables you can change.
Dependent variables you cannot change... they occur as a result of changes to independent variables.

If your data assumes a variable is dependent -- you cannot use that data to predict what would happen when that variable becomes the independent variable.


...
...
((Unless your data is a one-to-one mapping like y=mx+b... however y = x^2 for instance... you cannot say y = 4 so x MUST be 2... because x could be -2))

Yes... the math still works.

((Good question though))

[rossmum]

I think it's not so much they don't understand it as they're just so blindly pig-ignorant that they'll skew anything they can in their favour and call it legit

[Ifafudafi]

UUUUGGGGHHH ALGEBRA MAKE IT STOOOOOP

Essentially, from what my pathetic literature-oriented mind can gather, the problem here is that they're assuming that anything a violent person does can directly lead to the person becoming violent, as opposed to several risk factors (poor mental health, abusive childhood, and maybe violent vidya) converging to give us one psychotic murderer. Parents can sue you for slander, kids are pure and innocent, but video games can much more easily be marked as evil, because anyone with the power to properly defend them is too old to know anything about them. The perfect scapegoat.

[sdavis117]

Most serial killers eat carrots.

Carrots cause violence.

[n00b1n8R]

All murderers drank water.

Water leads to murdering guyz

[Bodzilla]

i think i've had too much to drink

having trouble following ya Babe too much lingo jingo mumble gumbo

[MastaCheefa]

Kind of like the 'all heroin addicts have used marijuana thus marijuana is a gateway drug to heroin' argument. But all peopel that use marijuana dont go on to use heroin. Yet you can bet most heroin addicts have tried everything out there.

People will twist numbers in any way they can if it will favor theyre agenda.

[Phopojijo]

Here's the problem... a lot of them are convinced they are actually right. If they were jerks, okay -- whatever. But some of these people think that their Junior High School math-violating knowledge of statistics is right.

[Mass]

78% of statistics are made up anyway

I don't think that people should use a lone statistic as evidence of anything, rather a variety of statistics should be used by a group of skilled researchers to reach a conclusion and the conclusion reached by that study ought to be the evidence, not any single datum they might have happened to acquire.

Data has to be evaluated and counterbalanced against any number of things that could skew it. Ultimately, I think that trends ought to be compared and examined with far more than two variables, and the simplification necessary for publication screws up the facts. The public has little usable training for looking at statistics and is highly likely to get the wrong impression from looking at one, (often enough, this is the point) the opinion of the researcher (obviously we hope this to be as uncolored as possible) is of significantly more value.

I'm going to note here that, because of human nature, despite this opinion of mine, I think all data should be viewable at any rate, so that holes (intentional or not) can be spotted.

[kid908]

100% of statistics are bias. I really hate statistics, but also find it funny how much Americans struggle with math (Of course I don't have 100% understanding of math, but who does?) and English... yeah they need to learn their own language.

They DO NOT need to "learn math." If you make up the numbers, you really don't need math since it's wrong from the start.

[jcap]

This is exactly the same soapbox I was on a few months ago when I gave a lecture to an entire class about twisting statistics.

The example I used was about an article saying that many people who developed prostate cancer masturbated when they were younger. However, a few months before (and over the past several years) there have been several studies saying that masturbation dramatically reduces the rick of prostate cancer.

It turned out it was a private company that funded and conducted the study of a small audience, but that's just a little extra thing to know... Anyway, the problem was that they isolated it to already a small group. The correct way to do it is the second way: taking a LARGE random selection, asking them if they masturbated when younger, and then asking them if they developed prostate cancer.

Also, GREAT algebra example.