Discussion: View Thread

  • 1.  Large F ration with insignificant p-value

    Posted 10-26-2011 16:25

    I had a student bring by an OLS regression printout that included an F value in the range of 7345 with a corresponding p-value of 1.0. I am unfamiliar with the specific software though it is used by Geography. I've never encountered an F statistic that large with a corresponding non-significant p-value. Any one ever experienced this type of output, and further, can it be anything other than an error?

    I sent her to extract the data so we can run it on software I'm familiar with to check the results.

    -------------------------------------------
    Byron Davis
    Staff Consultant
    Univ of Utah
    -------------------------------------------


  • 2.  RE:Large F ration with insignificant p-value

    Posted 10-26-2011 16:30
    Whatever the numerator and denominator degrees of freedom are it is wrong.  Either the value of the F statistic is wrong or the p-value should be something like p<0.0001.

    -------------------------------------------
    Michael Chernick
    Director of Biostatistical Services
    Lankenau Institute for Medical Research
    -------------------------------------------








  • 3.  RE:Large F ration with insignificant p-value

    Posted 10-28-2011 18:52
    I am assuming the software is approximating the denominator degrees of freedom and that approximation is providing something close to 0 or negative.

    -------------------------------------------
    George Milliken
    -------------------------------------------