SURVEY RESULTS: Axis or Allied Preferences

[Fionn]

Actually Mr. Peng, when a statistical probability is P 0.05 in medicine then it is assumed that the result is not due to chance.

Using probability to determine absolute certainty ... Hmm, there's something for the philosophers to ponder

------------------

___________

Fionn Kelly

Manager of Historical Research,

The Gamers Net - Gaming for Gamers

[Guest aaronb]

Philosophers and quantum physicists.

I prefer the Axis because, in my youth, I was forced by by gaming friends to take the Russians while playing PanzerBlitz and ASL. Now I can use the 'cool' toys (Tigers).

Fionn: I'll resend our latest turn shortly. Things have been hectic... oh, and I'll offer a cease-fire in three turns regardless of where we are...

[MrPeng]

While I understand the concept of statistical signficance, I still must take issue with the word "proves." Yes, while an alpha level of less than 5% is considered to be beyond random chance, it still isn't really a good idea to say that you have PROVEN something. It is a semantical argument and one of style not of fact. There are not many journal articles I've read that sum things up with, "Our research PROVES that daily beatings, ritual animal abuse and smoking before the age of 7 make wargamers want to play on the side of the Soviets." One would typically see, "Our research INDICATES that forced enemas, nipple piercing and being from New Jersey make wargamers post long goofy messages arguing over the appropriate use of a single word."

Respctfully submitted,

Peng - splitting hairs -

[Fionn]

Well, in medical journals a P of <0.05 is taken to be that level at which a causal relationship can be established between various factors and also that level where the word "proof" can begin to be used.

In true statistical probability studies virtually nothing is ever truly proven. Every discipline has to have a cutoff point where a low probability of this being chance = proof that this is not chance. Medicine just sticks with a P < 0.05

Other disciplines would vary I'm sure.

------------------

___________

Fionn Kelly

Manager of Historical Research,

The Gamers Net - Gaming for Gamers

[Bastables]

Sniff, I love you guys, all of you.

(At the behest of Fionn it is now guys as opposed to my good mate guy )

[This message has been edited by Bastables (edited 01-13-2000).]

[Fionn]

Bastables, who's Guy?

(obviously meant to be guys but ya kanna pass oop a chance like dat ya ken? )

------------------

___________

Fionn Kelly

Manager of Historical Research,

The Gamers Net - Gaming for Gamers

[Neutral Party]

The level chosen for statistical significance (0.05, 0.01, 0.005 etc,.) is the likelihood that an observed relationship could have happened by random chance. For a p level of 0.05 there is a 5% chance that the relationship could have occured randomly. There is no causal relationship implied by these statistical associations. This requires an interpretation of the results that statistics alone cannot provide.

Joe

[Fionn]

Ah,

Don't get an Irishman started on something relating to theology (reference to the whole interpretation of facts vs immutability of facts) . We were born to talk.

Neutral party, as you say interpretation of the statistical probability is necessary but the presence of interpretation does not negate, IMO, the validity of a result..

In life NOTHING is certain yet the word certain is used in relation to many things. This suggest that, heuristically speaking, the meaning of the word is not the same as the definition of the word (again interpretation and accepted usage come into the picture).

Finally, the whole issue of reality being subjective and tinted by one's own experiences as they influence ones outlook and thus interpretation of external, and indeed internal, events.

However, strictly speaking you're right. I would, however, aver that "in common usage" I am .

Ah, redefinition of terms , the last refuge for compromise in discussion hehe.

------------------

___________

Fionn Kelly

Manager of Historical Research,

The Gamers Net - Gaming for Gamers

[Neutral Party]

All righty then. Strictly speaking you're wrong but "in common usage" you're right. We can therefore interpret "in common usage" to mean "things people say who think they are right but are in fact wrong."

Unless of course we redefine right to mean wrong in which case you're completely correct.

Joe

[MrPeng]

By Golly, I think NeutralParty is absolutley correctamundo in his analysis of the rightness of being wrong. I am, however completely certain that we have Proven the fact that there can be more than one way to skin a sadistic statistic. Given a large enough sample size one can get a significant result even if the effect of the factor under study is not very robust at smaller sample sizes. Damn! i forget the name of the test for effect sizes. ah well, if there is still some interest in this thread later on i may come back with a more thoroughly researched riposte. Damn me, damn me to hell.

Peng, out (of alcohol)

[Fionn]

Neutral Party,

I KNEW the irresistible weight and self-evident correctness and moral rectitude of my overpowering argument would turn you to the dark side... umm.. I mean, you'd see my point ...

Seriously though, strictly speaking you're right of course but having two parents who write books on heuristics and inter-relatedness of phonemes etc etc gives me the outlook that not only can 2+2=5 but the word CAN equal the word wrong so long as you define your terms right (or should that be wrong LOL )

------------------

___________

Fionn Kelly

Manager of Historical Research,

The Gamers Net - Gaming for Gamers