PDA

View Full Version : An Introduction to Statistics and Research Methodology



pick6
12-15-2011, 06:57 PM
I am of the opinion that statistics alone can not be used as the basis for any theory or hypothesis. This is especially true when discussing matters of football, this being so do to the fact that football is best deliberated anecdotally. With that said statistics can play a role in proving your point but only in a secondary or tertiary capacity.

This is because nearly anything can be made to seem more valid with the addition of statistical information properly extracted and placed. As evidence, I will propose an obvious absurdity to demonstrate, albeit with an extreme example, how statistics can be used incorrectly.

Murder is caused by ice cream. Ridiculous, of course. But no, I have proof in the form of the almighty statistical data. Statistics clearly show that as ice cream sales increase so do the incidents of murder, across time and place. So in America if plotted data across time for both murder and ice cream sales we will see that they move together. The would be statistician might observe this correlation and apply for funding to look deeper into this amazing phenomenon. But the average Steeler fan would simply say "slow your roll, Sherlock" this seeming correlation can be explained in other ways. Perhaps its the hotter, longer, summer days or the fact that more people are outdoors or even out of school (among young people). Who knows, but it darn sure ain't ice cream, Einstein.

So in football terms, don't say that QB X is better than QB Y because QB X won more superbowls. You have built an argument on a false premise. Trent Dilfer is not better than Dan Marino, c'mon. But according to the above statistic of superbowl wins then this must also be true.

I know I'm oversimplifying, but my point is that all too often people shoot around stats as the key to their position when sometimes you might need a little more to make the point. Be careful with the stats guys they may be just part of the picture.

pittpete
12-15-2011, 07:05 PM
Too many big words, please oversimplify :P :P :P :P

RuthlessBurgher
12-15-2011, 07:27 PM
Too many big words, please oversimplify :P :P :P :P

Words good. Numbers bad. :D

Dee Dub
12-15-2011, 09:30 PM
Or how about this one? Just because said QB has thrown for a 400 yard game or has 3,500 plus yars in his first 13 games in the NFL doesnt mean he is an elite passer. :wink:


:lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol:

pick6
12-15-2011, 10:54 PM
Too many big words, please oversimplify :P :P :P :P

Words good. Numbers bad. :D


Thanks Ruthless, a brief mention of context and this is really all I had to say.

pick6
12-15-2011, 10:58 PM
Or how about this one? Just because said QB has thrown for a 400 yard game or has 3,500 plus yars in his first 13 games in the NFL doesnt mean he is an elite passer. :wink:


:lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol:


Im surprised anybody was able to stay awake through my post, but yeah you speak the truth.

snarky
12-15-2011, 10:59 PM
Or how about this one? Just because said QB has thrown for a 400 yard game or has 3,500 plus yars in his first 13 games in the NFL doesnt mean he is an elite passer. :wink:


:lol: :lol: :lol: :lol: :lol: :lol: :lol: :lol:

Nor does it mean he isn't.

snarky
12-15-2011, 11:04 PM
There's stats and there's statistics and very little (if anything) on this board that utilizes the latter.

PS: Correlation != Causation

pick6
12-15-2011, 11:14 PM
There's stats and there's statistics and very little (if anything) on this board that utilizes the latter.

PS: Correlation != Causation

I'm not familiar with the notation != but if that means correlation NEVER means causation then you are correct. But if it means that correlation = causation then this is incorrect. I think.

snarky
12-15-2011, 11:16 PM
There's stats and there's statistics and very little (if anything) on this board that utilizes the latter.

PS: Correlation != Causation

I'm not familiar with the notation != but if that means correlation NEVER means causation then you are correct. But if it means that correlation = causation then this is incorrect. I think.

! is used to represent not so != means not equal.

pick6
12-15-2011, 11:19 PM
There's stats and there's statistics and very little (if anything) on this board that utilizes the latter.

PS: Correlation != Causation

I'm not familiar with the notation != but if that means correlation NEVER means causation then you are correct. But if it means that correlation = causation then this is incorrect. I think.

! is used to represent not so != means not equal.

Thanks, I learned something.

Shoe
12-16-2011, 12:42 AM
I am of the opinion that statistics alone can not be used as the basis for any theory or hypothesis. This is especially true when discussing matters of football, this being so do to the fact that football is best deliberated anecdotally. With that said statistics can play a role in proving your point but only in a secondary or tertiary capacity.

This is because nearly anything can be made to seem more valid with the addition of statistical information properly extracted and placed. As evidence, I will propose an obvious absurdity to demonstrate, albeit with an extreme example, how statistics can be used incorrectly.

Murder is caused by ice cream. Ridiculous, of course. But no, I have proof in the form of the almighty statistical data. Statistics clearly show that as ice cream sales increase so do the incidents of murder, across time and place. So in America if plotted data across time for both murder and ice cream sales we will see that they move together. The would be statistician might observe this correlation and apply for funding to look deeper into this amazing phenomenon. But the average Steeler fan would simply say "slow your roll, Sherlock" this seeming correlation can be explained in other ways. Perhaps its the hotter, longer, summer days or the fact that more people are outdoors or even out of school (among young people). Who knows, but it darn sure ain't ice cream, Einstein.

So in football terms, don't say that QB X is better than QB Y because QB X won more superbowls. You have built an argument on a false premise. Trent Dilfer is not better than Dan Marino, c'mon. But according to the above statistic of superbowl wins then this must also be true.

I know I'm oversimplifying, but my point is that all too often people shoot around stats as the key to their position when sometimes you might need a little more to make the point. Be careful with the stats guys they may be just part of the picture.

You're correct... and I like this kind of stuff.

You mentioned Dan Marino. I do think it's relevant, when considered the GOAT, to base his inclusion in the conversation on the fact that he never won one. It's not statistical or based on any firm footing, but I just don't think you can be the GOAT if you never won a championship. You don't have to have won multiples, but you have to have had one..

Good post nonetheless.

Chadman
12-16-2011, 12:46 AM
Wait....so winning SB's doth not maketh the QB great?

Does that mean Ben sucks??


:D