Results 1 to 2 of 2

Math Help - One-group pretest-posttest analysis?

  1. #1
    Newbie
    Joined
    Apr 2010
    Posts
    1

    One-group pretest-posttest analysis?

    Hello all,

    Experimental design:
    8 subjects, each run through 6 identical trials in a pre-test. Measures were collected from each trial.
    Each of the subjects was then put through a treatment (exercise program).
    The 8 subjects then returned for another 6 identical trials for a post-test.

    To analyze this data, it would appear (from what I've been reading) that the most appropriate test would be a 2-sample paired T-test. Can anyone confirm this? I know that this is based on the assumption that the data is normally distributed - how do I go about checking this assumption? I'm using Minitab v15.

    Additionally, what do I do if it turns out the data is not normally distributed? Again, from what I've been reading, it would appear that I should attempt to transform the data, and that if this fails, I should use a non-parametric test (Wilcoxon in this case, I believe). Can anyone confirm this as well?

    Your speedy help on this matter would be greatly appreciated =)

    Cheers,
    Mario
    Follow Math Help Forum on Facebook and Google+

  2. #2
    Newbie
    Joined
    Apr 2010
    Posts
    3
    Paired-sample t-test is fine, but if you want to do any comparison within the 6 trials, you should use repeated-measure ANOVA. 8 participants are too few for normal-distribution test. Normally, if you don't have a reason to believe the data is not normal-distributed, you can just use parametric test.

    Quote Originally Posted by FoosJunkie View Post
    Hello all,

    Experimental design:
    8 subjects, each run through 6 identical trials in a pre-test. Measures were collected from each trial.
    Each of the subjects was then put through a treatment (exercise program).
    The 8 subjects then returned for another 6 identical trials for a post-test.

    To analyze this data, it would appear (from what I've been reading) that the most appropriate test would be a 2-sample paired T-test. Can anyone confirm this? I know that this is based on the assumption that the data is normally distributed - how do I go about checking this assumption? I'm using Minitab v15.

    Additionally, what do I do if it turns out the data is not normally distributed? Again, from what I've been reading, it would appear that I should attempt to transform the data, and that if this fails, I should use a non-parametric test (Wilcoxon in this case, I believe). Can anyone confirm this as well?

    Your speedy help on this matter would be greatly appreciated =)

    Cheers,
    Mario
    Follow Math Help Forum on Facebook and Google+

Similar Math Help Forum Discussions

  1. Order of Group. Direct Product of Cyclic Group
    Posted in the Advanced Algebra Forum
    Replies: 9
    Last Post: November 19th 2011, 01:06 PM
  2. Replies: 1
    Last Post: November 4th 2009, 09:52 AM
  3. Quick questions on Group Theory - Cosets / Normal Group
    Posted in the Advanced Algebra Forum
    Replies: 3
    Last Post: October 16th 2009, 08:39 AM
  4. Group Theory Question, Dihedral Group
    Posted in the Advanced Algebra Forum
    Replies: 0
    Last Post: March 4th 2008, 10:36 AM
  5. college pretest
    Posted in the Pre-Calculus Forum
    Replies: 4
    Last Post: June 26th 2006, 03:44 PM

Search Tags


/mathhelpforum @mathhelpforum