TP for summing up a lot of what I was going to say. I don't want to speak for anyone specifically but there seems to be an attitude, in and out of this thread, that "If one statistic is misleading, this proves that all statistics are misleading!" Ironically this belief displays a lack of statistical understanding, and often gets used to justify continuing to be that way.
Minny being 3rd in the league at turnover differential is only a "lie" if you think the stat argues that good turnover differential can only happen for good teams. Of course the stat doesn't state or imply any such thing. It just means an otherwise bad team seems to be good at one aspect of the game. You have to misunderstand or misinterpret what the statistic means to claim it's "deceptive" in some way.
People, on the other hand, will often wield statistics in a misleading way, but that's a different story, and yet another reason why an independent understanding is useful.
All stats are midleading... well maybe not the stats, but the person reporting the stats has the ability to make them say whatever they want to say. In that way stats are always misleading. I don't know what else to say... you have a mathmetician on here saying that stats lie... I mentioned my statistics professor who said stats lie, I've given a very valid argument... and no I don't have an attitude and yet you say I have an attitude and continue to downplay my reasons for saying stats lie. You simply can not believe all stats, that's all there is to it.
Nothing personal, but this is kind of a throw the baby out with the bathwater approach. Stats can be misinterpreted or misleading, but that doesn't mean that they can't be used properly or that they can't paint an accurate picture.
If you went back and read my very first post on the subject I said there are obviously some exceptions. All I am saying is that there are some cases where stats lie... I'm not saying they do in every situation, but people have to be aware that they're not always accurate. And no worries, I don't take any of this personally, I actually find this to be a very interesting discussion. I just wish people would admit that I am right that stats lie in some cases... because I know I am. Stats are not always accurate. TPs all around for a good discussion!
You're changing your stance though! I think clearly even the "stat advocates" have been saying all along that you have to be clear and careful. You were literally saying that stats are always misleading! that stats lie!
But it's not arbitrary. Just because one person can make up numbers to support their point does not mean that all statistical analyses are equally inept.
If I said "Stat's always lie" I misspoke, my bad. I did say stats are misleading and I mean that, and once again my very first post said there are cases when stats are truthful and useful, so no I'm not changing my stance, that's where it was in the beginning.
I'm saying stats work in certain situations but people alter statistics in order to influence other peoples views on the stats... that is just fact.
I agree, but that's a problem with people misusing and misinterpreting statistics, not with statistics themselves. I think this is where the two sides aren't quite connecting. A legitimate statistical finding can be represented accurately or dishonestly, and it can be interpreted accurately or incorrectly. In neither case does the stat itself change, just how someone is using or viewing it. Which is why I think it's important that people understand statistics - because they are so meaningful but also so easily abused and misread.
BTW, I only replied to one post of yours (technically VT's), I haven't been arguing with you throughout - also, I meant "attitude" in the generic sense of a belief, not in the "bad attitude" sense. It's absolutely no knock on you personally. Apologies for any misunderstanding.
And with the appeal to authority thing, I'm not a statistician or a mathematician, but I have a doctorate with a concentration in Quantitative Psychology, which basically means statistical analysis and methodology in psychological research. I use statistical analysis every day at work and teach introductory statistics courses as well. I know some of the other people who disagree with you have some formal training too.
Like I said I think this mostly boils down to whether you attribute things to the statistics themselves or the people who interact with them. It sounded for awhile like you were doing the former, which is why so many people disagreed. Looks like you may have meant the latter all along though. TP back at you for discussion.