Notices
Results 1 to 10 of 10

Thread: the value and distortion of statistics

  1. #1 the value and distortion of statistics 
    Forum Freshman
    Join Date
    Nov 2005
    Posts
    5
    deleted


    Reply With Quote  
     

  2.  
     

  3. #2  
    Forum Ph.D.
    Join Date
    Oct 2004
    Location
    Maastricht, Netherlands
    Posts
    861
    Though I'm not exactly full on ideas on it, I do applaud it. I'd be interesting in reading the entire essay, once you have finished it. Myself I am getting all too tired as well of the abuse of statistics by journalists who bend psychological, per example, research to produce sentences that are true, but to the general public are misinterpreted.

    Mr U


    Reply With Quote  
     

  4. #3  
    j
    j is offline
    Forum Bachelors Degree
    Join Date
    Mar 2005
    Posts
    431
    Quick point:

    The very questions asked effect the statistics generated. This doesn't refer to the unavoidable bias in language, but also the decision of what to measure.

    For example, on war casualties, does one measure deaths, deaths and maimings, or all injuries?
    Why do they want us to believe Conspiracy Theories?
    Reply With Quote  
     

  5. #4  
    Forum Ph.D.
    Join Date
    Oct 2004
    Location
    Maastricht, Netherlands
    Posts
    861
    Of course, but the definition, and the recognisition of that definition is vital in understanding the statistic. I would consider it interesting to include for a complete view.

    Mr U
    Reply With Quote  
     

  6. #5  
    Universal Mind John Galt's Avatar
    Join Date
    Jul 2005
    Posts
    14,168
    I'm sure you are familiar with the quote "Lies, Damned Lies, and Statistics"
    attributed to Disraeli or Mark Twain. [In checking on the authorship of that quote I ran across this interesting link, which describes an abominable use of statistics - this may be of use to you. http://www.ucpress.edu/books/pages/9358/9358.intro.html ]
    I can't agree with your assessment that statistics can be misleading because it is a simplified form of language. Quite the reverse. The strength of statistics lie in their simplicity, directness and, within context, their completeness. The problems with statistics arise from the following:
    1) Improper definition of terms.
    2) Improper definition of scope.
    3) Ignorance on the part of the end user of the importance of 1) and 2).

    Good luck with the project. As HU suggested, why not post your completed version here.

    Edited to add overlooked link.
    Reply With Quote  
     

  7. #6  
    Forum Freshman
    Join Date
    Nov 2005
    Posts
    5
    thanks for the suggestions and that link. It has all been very useful!
    Reply With Quote  
     

  8. #7  
    Him
    Him is offline
    Forum Sophomore Him's Avatar
    Join Date
    Aug 2005
    Posts
    181
    I always have been intrigued by statistics and large scale analysis. In DNA microarrays or proteomics study’s one studies (sometimes) over 1000 different expressions patterns, looking for different patterns in two situations. In scientific journals a p value under 0.05 is regarded as significant, however studying 1000 genes, you can expect 50 genes showing (supposable) significant differential expression (when comparing two situations) purely by chance.
    Maybe this is why pure DNA microarray and proteomics analysis are more and more refuted, and only excepted if confirmation experiments are employed.

    In summary, watch out for large scale analysis, you’ll easily find interesting differences, which look statisticly valuable.

    A second point is the bad knowledge of many (otherwise very qualified) researches have of statistics. The wrong use of statistic test is often the case. Must admit my knowledge of statistics could also be better; so i am not judging.


    And lastly the use of correction factors makes you can manipulate your results. Different logical approaches often lead to very different results. It asks much discipline of the researches not to choose the most attractive solution (this which confirms your hypothesis) over the most correct (this which explaining al the data). Often contradicting results are not reported and only the clean wrongly normalized data is used to prove certain point.

    In a new summary:
    Statistics has a lot of tricks to make your results fit your idea, and knowledge of which tricks are allowed is not always present.
    =dangerous.
    he who forgets...will be destined to remember (Nothing Man - Pearl Jam)
    Reply With Quote  
     

  9. #8  
    Forum Freshman
    Join Date
    Nov 2005
    Posts
    5
    deleted
    Reply With Quote  
     

  10. #9  
    Forum Ph.D.
    Join Date
    Oct 2004
    Location
    Maastricht, Netherlands
    Posts
    861
    Nice! A tip, keep the webaddresses, and other sources written down somewhere you won't lose them. Preferably written in the APA-norm, so you can immediately type them into your essay, with proper footnotes and everything. Sources are important, and they make your essay more professional.

    Not sure whether I'm mr obvious now, but that's what I always got wrong at my other study :P

    Mr U
    Reply With Quote  
     

  11. #10  
    Forum Bachelors Degree charles brough's Avatar
    Join Date
    Jun 2005
    Location
    joplin MO USA
    Posts
    425
    Maison, your request for information on the misleading aspects of statistics has brought out a lot of good information!

    May I just add a suggestion that you concentrate a while on the subject of rationalizing. Dig into the personal reasons and, as well, on common ways which the whole social science profession rationalizes in order to avoid conflict with religion and secular-political beliefs.

    In http://humanpurpose.simplenet.com there is an Appendix that lists 21 rationalizing tools I have dug up and dug out of the books and papers I have read in the social sciences.

    Also, the point was well made in the other posts that how words are used is important. It is common only to the social sciences that words have a lot of different meanings. This is essential to rationalizing.

    Another example from that shown above is the word "torture." NO one ever committs torture because whoever committs it defines it as more brutal than what they have done! Yes, the US government does torture terrorists if we define the word as the inflicting of suffering to get information or to punish.
    Brough,
    civilization-overview (dot) com

    --------------------
    There are no accidents, just someone taking too much risk. . . (CB)
    Reply With Quote  
     

Bookmarks
Bookmarks
Posting Permissions
  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •