ceigered:
Back to the original topic, what's better: statistics that you can't trust, or no information at all that you can't trust either?

Or as Mark Twain once quipped: there are three types of lies: lies, damned lies, and statistics.

Or this one by Ernest Rutherford, "If your experiment needs statistics, you ought to have done a better experiment. "

On the other hand, Gerd Gigerenzer in "Calculated Risks" reports that at the Napoleonic court, the saying was: "If you want something from Napoleon, give him statistics". Not to mention the chestnut from Benjamin Franklin: "In this world nothing is certain except death and taxes."

Doesn't the quote with relation to Napoleon suggest that he'll give you anything, whether you really need it or not, so long as you have some kind of number to back you up?

Anyway, people tend to parade around the statistics supporting what they believe and disregard statistics that contradict them.

erinja:..people tend to parade around the statistics supporting what they believe and disregard statistics that contradict them.

The fact that something can be misused does not show that it does not have a legitimate, even necessary role. I would recommend Gigerenzer's book (updated in 2003 as Reckoning with risk) as a valuable contribution to public education.

erinja:Doesn't the quote with relation to Napoleon suggest that he'll give you anything, whether you really need it or not, so long as you have some kind of number to back you up?

Anyway, people tend to parade around the statistics supporting what they believe and disregard statistics that contradict them.

This is very narrow minded way to look at things. Didn't you say before that every person should know another alphabet and I said that was arrogant when people didn't know basic science?

Well here you demonstrated a profound lack of insight into statistics. Statistics is not the pop-facts and figures you see in newspapers where it says 40% of blaa did blaa!!wooo

Statistics is an analytical tool which can very precisely tell whether some data which exhibits a pattern actually does exhibit a pattern. It can put interval ranges for measurements that you know are accurate 99% of the time. Those quotes you all said are used to illustrate a point- not to say statistics is bad you numbnuts!

Statistics = a tool. Tools can be (mis)used however you want.

Hahaha.

Hey, erinja, how did you ever get your engineering degree without learning anything about science or statistics?

Good catch, Genjix! Apparently you've found something missing from the US engineering curriculum!

----------
This is sarcasm, btw, for those who are more or less humor-challenged.

I would never say that statistics are bad, or that they aren't useful. But if someone gives you a number and asks you to believe what they say because of this number, then you're a fool if you believe them without asking for some more information.

Actually, I count statistics among the most useful math courses I ever took. In my current position, I never use any of the calculus or higher level math that I studied. But statistics are important to understand.

Obviously I don't really know what I'm talking about since my degrees are both in mechanical engineering and not statistical methods. It must be that the schools where I studied didn't teach me what I need to know; I took a semester of statistics in high school and another in my undergraduate university, and I know that this isn't a lot. I humbly admit that I didn't take even one course in statistics as part of my master's degree studies. Clearly I shall have to write to MIT and tell them to amend their graduate school curriculum to fill this hole.

Using my paltry two semesters of study in statistics, I made extensive use of statistics in my graduate thesis work. Therefore I am very well aware of their capabilities and limitations. Working with the data I collected in my thesis research, I ran a number of different statistical tests, and I found that my results were considered statistically significant or not significant depending on which test I chose. This result proved to me the ambiguity of some of my data. It also proved to me that it's possible for a dishonest person to take ambiguous data and package it in a way that gives a misleadingly positive (or negative) result.

So as an engineer I'm a strong believer in the power of statistics. And as a person who has spent most of her life in the suburbs of Washington DC, where national politics are local news, I'm acutely aware of the power of misuse of statistics. Suffice it to say that when I see a number that "proves" something, since I am an engineer and a cynic, I want to know (a) who did the testing that produced these numbers, and whether they had a vested interest in the outcome; (b) whether the full data set or only a partial data set was used to generate the statistical results; (c) whether all factors were adequately controlled for in the experiment, or whether an unrelated factor may have caused this result; and (d) whether the statistical tests that were done on the data were really appropriate to the type of data that had been collected.

If you believe a number without asking these questions, you are begging to be misled by someone dishonest.

Right, gotcha. I misunderstood you earlier as just attacking all of statistics.

What you refer to about blindly being led by the nose, I think of as a lack of critical thinking. Many times I see attacks on sites like Wikipedia because zomg!? it can be edited by anyone!!11! Articles on Wikipedia are far more accurate than the junk in general press/media. You see things quotes as absolute truth, go check them on the net and find that well... actually it's under heavy debate or an old viewpoint.

Because of the internet (and the ability from it to validate things I read from many sources), I've become highly skeptical of surprising information I read. I assign green/yellow/red from highest to lowest for trustworthiness of info I read. Untrusty sources are red, trusted or many confirming sources get green. Sadly most general media falls on red or yellow. Compare this with a person like my father who gets all his knowledge from TV and believe all kinds of stupid ideas.

I used to be a full-time hacker, and whenever I read articles on tech in newspapers it just makes my squirm how terrible the info is. And yet somehow we imagine they understand all the other issues they report on. I'm glad that we have the internet now with tools like Wikileaks, Wikipedia .et al for educating ourselves. For being skeptical reasoners. That's how creative independent thinkers are produced- in the marketplace of ideas.

RiotNrrd:This is sarcasm, btw, for those who are more or less humor-challenged.

I used to use statistics a lot. Easy to fool yourself if you want to believe something. However if you choose to accept the result then you can get unambiguous answers.

Didn't know she was an engineer I misremembered her as a language teacher for unknown reasons. I was wrong. Don't take it as an attack- my tone was trying to be playful