Discussion: View Thread

  • 1.  RE: Correlation between X and Y, where each Y has multiple X values

    Posted 04-26-2011 18:24

    Personally, I don't mind seeing this thread -- I find it educational.

    For those who prefer, there is a means to eliminate the emails, or to receive them as a daily digest.  Click on "eGroup Home" and follow the links to the personal profile section.

    -------------------------------------------
    Thomas Nunnikhoven
    International Paper
    -------------------------------------------


  • 2.  RE:Correlation between X and Y, where each Y has multiple X values

    Posted 04-27-2011 10:31

    I think it is very interesting and appreciate the sharing and exchanges.-------------------------------------------
    J. Dobbins
    Delmarva Foudation
    -------------------------------------------








  • 3.  RE:Correlation between X and Y, where each Y has multiple X values

    Posted 04-27-2011 11:59
    I'll put in one last 2 cents here.

    From my looking on the internet, the hierarchical model approach specifically addresses the problem that Beverly put before this list.serve, but is based on a Bayesian model and is, from what I looked at, difficult to implement.  The measurement error in X could be important, but, for simple linear regression, an easily implemented approach to taking into account the error in X can be found in the work of Wayne Fuller (who was my major professor, although I never worked with him on measurement error theory), given that there is available a good estimator of the variance of X, which Beverly has.  The presence of measurement error in X biases down the estimate for beta when simple regression is done.  However, according to Fuller, even in the presence of measurement error in X, for simple regression, the t test for beta equal to zero is valid, just not as powerful as the test could be.  

    Margot
    -------------------------------------------
    Margot Tollefson
    Owner
    Vanward Statistical Consulting
    -------------------------------------------








  • 4.  RE:Correlation between X and Y, where each Y has multiple X values

    Posted 04-28-2011 13:00

    For those of you who work in groups with other statisticians, this may seem a bit protracted to you, and perhaps boring.  However, to us singletons who have little dialog with other statisticians, this has been most entertaining and informative.
    -------------------------------------------
    Richard Browne
    Texas Scottish Rite Hospital for Children
    -------------------------------------------