coff of correlation r

Discussion in 'CT3' started by gcpgcp, May 13, 2009.

  1. gcpgcp

    gcpgcp Member

    Nov 2008 exam Act Soc India

    Q. 17) As humidity influences evaporation, the solvent balance of water-reducible paints during spray-out is affected by humidity. A study is conducted to examine the relation between humidity (X) and the extent of solvent evaporation (Y). The following data summary is obtained:
    n = 25, Σx = 1314.90 Σy = 235.70, Σx2 = 76,308.53, Σy2 = 2286.07 Σxy = 11,824.44
    a) Find the estimated correlation coefficient between X and Y and test the
    hypothesis H0 : ρ = 0, against H1 : ρ≠ 0 ; ρ being the population correlation
    coefficient between X and Y.


    ANSWER given at the ASI website is: r = -0.847
    For testing the hypothesis H0 : ρ = 0 against H1 : ρ ≠ 0, the test statistic is
    t= r (n-2)^1/2 / (1-r^2)^1/2 = -7.635
    The table value of t23,0.05 is 1.714. Reject H0

    MY QUESTION:

    ISNT IT A 2 SIDED TEST ? IF YES, WHY t23, 0.05 is used ? IT SHOULD BE t23, 0.025 !!! assuming we are doing a 5% test.
     
    Last edited by a moderator: May 13, 2009
    Chandrima likes this.
  2. gcpgcp

    gcpgcp Member

    :confused:

    anyone ???
     
  3. nageshmcl

    nageshmcl Member

    No change in the decision.

    Since you are rejecting H0 at 95%, so by default you will reject H0 at 97.5% also.
     
  4. Chandrima

    Chandrima Member

    Even if we do the testing at 2.5% at both the sides, the also our conclusion will be right. And I feel, no marks should be deducted for that.
     
  5. John Lee

    John Lee ActEd Tutor Staff Member

    The "level" is the term for the rejection region - so you should say reject at the 5% level. And rejecting at the 5% level doesn't mean that you automatically reject at the 2.5% level. But the reverse is true.
     
    Chandrima likes this.

Share This Page