Tuesday, March 29, 2011

Statistics Questions for Test

If you have any questions regarding the test. Feel free to ask them here. I will post my responses to the blog.

A reminder the test will be in the lecture venue. You may bring one handwritten sheet of notes. Remember to bring a calculator.

Mike G

6 comments:

  1. I Have received this question

    When calculating Cohen's D for the upper and lower CLs, should we use the corresponding upper and lower CL for the Std Dev or do we always use the mean std dev?


    You always use then the "mean" actual std deviation.

    ReplyDelete
  2. In the 2009 honours test, question 2c asks us to calculate effect size for all 3 pairwise comparisons. I'm not certain what this is referring to as aren't there 6 possible pairwise comparisons?

    ReplyDelete
  3. In a Kolmorgorov-Smirnov test for skewness, which indicates that the distribution is normal - rejecting or failing to reject the null hypothesis?

    ReplyDelete
  4. The KS is more strictly a test of normality. The Null hypothesis is that the distribution is normal. As such we would ideally prefer to fail to reject the null hypothesis. The KS is not part of the test.

    ReplyDelete
  5. Am I correct in saing that for a 1 way ANOVA the R-square value is the global effect size?
    Also in a 1 way ANOVA, to caluclate the effect size for each pairwise comparison do we use the formula xbar1 minus xbar2 divided by the Root MSE?

    ReplyDelete
  6. Yes- R-Square = Eta Squared = Global Effect Size in 1 way anova = SS_Between/SS_Total

    Correct
    D = (xbar1 - xbar2)
    _______________
    RootMse

    ReplyDelete