1. Welcome to the All Things Male Forum. Please sign up and join the conversation. It's free!

How to properly interpret statistic data in medical articles

Discussion in 'Diet & Nutrition' started by dwx, Nov 8, 2012.

  1. dwx

    dwx Banned

    Joined:
    Mar 7, 2012
    Messages:
    167
    Likes Received:
    1
    I've never been good at math and am not sure what those small P numbers mean and what P stands for.

     
  2. Peabody

    Peabody Member

    Joined:
    Aug 7, 2012
    Messages:
    56
    Likes Received:
    1
    Most of the people who do these studies don't understand that either.

    But what I want to know is what your quote was from. Whatever is being described - I think I want some of that.
     
  3. hebsie

    hebsie Super Moderator Staff Member Super Moderator

    Joined:
    Apr 13, 2008
    Messages:
    3,198
    Likes Received:
    32
  4. bgnb

    bgnb Active Member

    Joined:
    Dec 9, 2007
    Messages:
    1,649
    Likes Received:
    19
    Here's the simple version: say you think that A can have an effect on B ... so you set
    up an experiment to test that hypothesis. Conventional research says that what you want
    to test is the Null Hypothesis and the P value is the point at which you ARBITRARILY decide
    that you would 'reject' the Null Hypothesis in favour of the Alternate Hypothesis.

    So one Null Hypothesis is that there is no effect, or that the effect will be more or less
    than a pre-determined value. You would then run the experiment, collect the data and
    analyse it using statistical methods.

    What you are looking for is this: Is this effect, if it exists, due to chance or something else?

    If the p-value is greater than the preset amount then you would basically fail to reject the
    Null Hypothesis ... if it is smaller, then you are basically saying that the 'chance of this effect
    occurring, if it exists, is less than simple chance' and you would reject the Null Hypothesis of
    no effect' in favour of the Alternate Hypothesis that says that if there is an effect, the likelihood
    of it occurring is less than chance.

    Where many 'researchers' mess up is the belief that a smaller p-value is somehow more significant
    when in fact it is simply less likely to occur by chance but still only indicates you can reject the
    Null Hypothesis.

    Without knowing what the research is about, what the previous research found, what the effect size
    might be and the sample size (which directly affects the p-value) then most of those P-values are
    meaningless garbage.
     
  5. Dr. John Crisler

    Dr. John Crisler Lord of the Forum Staff Member Super Moderator

    Joined:
    Dec 9, 2007
    Messages:
    10,758
    Likes Received:
    92
    Bottom line: it's the likelihood the results were not from mere chance.
     
  6. BadassBlues

    BadassBlues Super Moderator Staff Member Super Moderator

    Joined:
    Sep 2, 2011
    Messages:
    3,809
    Likes Received:
    123
    Its important to realize that studies are not in themselves the holy grail. They are important in their validation or invalidation of an assertion or theory. It's just another step in a very long and tedious process.
     

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.