iSixSigma

high p-value of constant in regression

Six Sigma – iSixSigma › Forums › Old Forums › Europe › high p-value of constant in regression

Viewing 3 posts - 1 through 3 (of 3 total)
  • Author
    Posts
  • May 26, 2009 at 8:30 pm #24116

    lost123
    ★★★★★★★★★★
    Participant

    When i run a multiple linear regression, i get very low p-values for most of the indipendent variables and for the model as a whole. BUT i get get very high p-values for the constant. I read in this forum what causes this. Now my question is: How does it influence the results of my tests and How can it be corrected?
    Thanks for any help,
    N

    0
    May 28, 2009 at 3:15 pm #57752

    Allard Munters
    ★★★★★★★★★★
    Participant

    Which software are you using? If you’re using Minitab, there is a possibility (via the options button), to remove the constant from the model by unchecking “fit intercept”.
    Allard

    0
    June 10, 2009 at 9:19 am #57753

    Remi
    ★★★★★★★★★★
    Participant

    Hai lost123,
    remember: for predictions of Regression it is recommended to only interpolate (i.e. only make predictions in the area where you have data).
    The high p-value of the constant says nothing else than that your prediction around the origin will be ‘bad’.
    You can correct this by extending your data collection to the area around the origin.
    Remi

    0
  • Author
    Posts
Viewing 3 posts - 1 through 3 (of 3 total)

The forum ‘Europe’ is closed to new topics and replies.