User User name Password  
   
Tuesday 2.12.2008 / 07:32 PM
Search:        In English   Suomeksi   På svenska
afterdawn.com / profiles / deepskystarlight / blog archive / statistics /
Home Blog Pictures Shoutbox Links

statistics

07 Apr 2007 0:26 (Edited: 07 Apr 2007 0:26)

Ok I've been taking this statistics class and it's really getting to me so I wrote this and posted this on our class site a couple of days ago and have gotten great responses to it. Hahaha you have to deal with the stress somehow.



It seems to me that some of the terminology of statistics is needlessly confusing. It’s one of those things that has always bothered me about math, in general. Math people love to break all the rules of clarity. The double and triple negatives are not generally used in English and writing manuals recommend against doing so, for clarity. Yet here you go:

The level of significance is used to indicate the probability that we are wrong in
rejecting the null. Also called level of probability, or p level, it is expressed as a decimal
and tells us how many times out of a 100 or 1,000, we would be wrong in rejecting
the null assuming the null is true. (In other words, how often we would expect no real
difference even though we rejected the null.)

Ok so let’s analyze this in English. The null hypothesis is itself a negative. It says that there is no effect, or no correlation, that the solution you are testing does not work. So that’s one negative. The level of significance is itself a triple negative because it is saying that the probability is that you are wrong (one negative) in rejecting (another negative) the null (a third negative). And people wonder why we have a hard time getting this. I am still trying to reword this into a clearer statement. I keep thinking that if I eliminate enough double negatives I can make a positive statement but it’s very confusing. You’d think I’d have a better chance of getting this since I know Spanish and French too but it’s too convoluted.

I will try to put this in English. Suppose my hypothesis is that statistics drive people crazy. The null hypothesis would be that it does not drive people crazy. Rejecting this hypothesis would mean that it does, in fact drive people crazy. (Or in statisticsese: You are saying it is not true that it does not drive people crazy). The level of significance is measuring the probability that if you reject the null hypothesis, that you would be wrong (statisticsese: you would be mistaken in saying that it is not true that it does not drive people crazy) --that in fact, it did not drive people crazy. So in English, you are saying that this is the probability that in saying it does drive people crazy you would be wrong because it doesn’t drive them crazy, but the very fact that statistics were involved means that we were driven crazy.

Yes, I realize that in statistics you never actually prove your hypothesis; you just prove that the statement that your hypothesis is false is wrong. But there has to be a better way of wording all of this.

 

List of my blog entries

User comments

    (No comments made)


Post your comment

In order to post your comments here, you need be logged in to our system. Simply follow this link in order to login and to post your comments here.

Digital video: AfterDawn.com | AfterDawn Forums | DVD X Copy Forums
Music: MP3Lizard.com
Gaming: Blasteroids.com | Blasteroids Forums
Software: Software downloads
Blogs: User profile pages
RSS feeds: AfterDawn.com News | Software updates | AfterDawn Forums
International: AfterDawn in Finnish | AfterDawn in Swedish | download.fi | fin.MP3Lizard.com
Navigate: Search | Site map
About us: About AfterDawn Ltd | Advertise on our sites | Rules, Restrictions, Legal disclaimer & Privacy policy
Contact us: Send feedback | Contact our media sales team
 
  © 1999-2008 by AfterDawn Ltd.