Re: Which test is appropriate?

From: DJ (droberts_at_midsouth.rr.com)
Date: 07/26/04

  • Next message: Austin: "Comparing differences in size of effect" 
    Date: 25 Jul 2004 19:59:11 -0700

    > Right -- standard for dependent samples, of which before-and-after
    > studies are a common example. The (signed) differences are the
    > actual data set: compute the mean and s.d. of the N differences and
    > run a standard t-test.
    >
    > Let me give you a simple illustration. You have twenty participants.
    > Each scores 99 on the dexterity test. After alcohol, each scores 98.
    > Comparing group means tells you nothing; but if every single person
    > lost one unit of dexterity you can be pretty sure that _something_
    > is going on, even if it's only persistent measurement error.

    Thanks for your insight and the explanation. Makes more sense the way
    you described it.

    DJ

  • Next message: Austin: "Comparing differences in size of effect"

    Relevant Pages