Closed Thread
Results 1 to 6 of 6

Thread: How do you find a P-Value from a T-Test?

  1. #1
    Points: 8,325, Level: 61
    Level completed: 59%, Points required for next Level: 125

    Posts
    1
    Thanks
    0
    Thanked 0 Times in 0 Posts

    How do you find a P-Value from a T-Test?




    Hi, I'm a student at Virginia Tech taking Biostat.

    With a T-Test, how do I find the P-Value using a t-table (df) and a z-table (z)?

    I have an exam tomorrow so help would be greatly appreciated!

  2. #2
    TS Contributor
    Points: 9,136, Level: 64
    Level completed: 29%, Points required for next Level: 214

    Posts
    889
    Thanks
    0
    Thanked 93 Times in 92 Posts
    To find the p-value when using a t-table:
    1) Look in the table for the row that matches your df
    2) Find which column of data your statistic falls between -> that gives you your range for the p-value.

    ex. df=7, statistic t=2.54 -> 0.01 < p-value < 0.025
    df=7, statistic t=5.18 -> p-value < .005

    To find the p-value when using a z-table:
    1) Using your statistic z, find the row that matches z up to the tenths place
    2) Staying in that row, find the column that matches z in the hundredths place -> the entry you're at can be used to find the p-value. You may have to do 1 - that entry, 2*that entry, etc depending on the test.

    ex. an upper-tail test, statistic z=0.96
    Look for row 0.9, then look for column .06 -> entry = .1685. Since the test is an upper-tail form, p-value = .1685
    ex. a double-tailed test, statistic z=2.42
    Look for row 2.4, then look for column .02 -> entry = .0078. Since the test is double-tailed, p-value = 2*.0078 = .0156

  3. #3
    Points: 2,041, Level: 27
    Level completed: 28%, Points required for next Level: 109

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Mean Joe, I believe CT is asking how to find the p-value within the range of 0.01 < p-value < 0.025 ... Does anyone know how to do that?

    Also, for df=7, statistic t=5.18, isn't the p-value in the range: 0.001 < p-value < 0.0005 (one-sided)?
    Last edited by Gina Gina Gina; 10-16-2009 at 03:39 AM.

  4. #4
    Points: 2,041, Level: 27
    Level completed: 28%, Points required for next Level: 109

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts
    Yeah, thank you stats TAs...

    There is no calculation or method for finding an exact p-value within a range of p-values (as determined by a t-table). You have to use a computer to determine an exact p-value.

    The p-values within the range do not follow a linear trend, so it doesn't really work to try to use proportions either.

  5. #5
    Points: 3,610, Level: 37
    Level completed: 74%, Points required for next Level: 40

    Location
    Mumbai, India
    Posts
    142
    Thanks
    0
    Thanked 0 Times in 0 Posts
    P values can be found only using Z tables.t tables are available only for certain given p values .01 ,.025,.05.0.1 etc.To find exact p value for t table resort to Excel functions.

  6. #6
    R purist
    Points: 21,251, Level: 91
    Level completed: 81%, Points required for next Level: 99
    TheEcologist's Avatar
    Location
    The Netherlands.
    Posts
    1,704
    Thanks
    229
    Thanked 451 Times in 258 Posts

    Quote Originally Posted by Gina Gina Gina View Post
    There is no calculation or method for finding an exact p-value within a range of p-values (as determined by a t-table).
    P-values are the 'area' under the curve greater than your t-score/z-score (depending on the tail). So calculation of exact p-values simply requires integration of the distribution under study. Often this only requires looking up the cumulative probability distribution.

    http://en.wikipedia.org/wiki/Cumulat...ution_function

    for the normal distribution look it up here;
    http://en.wikipedia.org/wiki/Normal_...ution_function

    for the t;
    http://en.wikipedia.org/wiki/Student%27s_t-distribution

    So there is a way for finding exact p-values, its well known and as old as the tests themselves. Its just that most textbook authors feel that this is too complex for most people so they calculated them for you and presented it in table form.

    FYI: Here's a free & simple program from the Uni of Groningen (Netherlands) that can calculate exact p-values for you;

    http://www.eco.rug.nl/~knypstra/Pqrs.html
    The true ideals of great philosophies always seem to get lost somewhere along the road..

Closed Thread

           




Similar Threads

  1. Find the P-Value. Use it to test the claim.
    By Spartan2014 in forum Statistics
    Replies: 3
    Last Post: 04-10-2011, 09:24 PM
  2. Replies: 1
    Last Post: 12-09-2010, 01:37 AM
  3. Cant find the right test for my dissertation. Please Help!!
    By clued_up_arsehole in forum Statistics
    Replies: 4
    Last Post: 10-18-2007, 02:33 PM
  4. Find the P-value for the indicated hypothesis test.
    By combat75 in forum Statistics
    Replies: 0
    Last Post: 09-23-2007, 02:50 PM
  5. Newbie Here! Question on find number of test...
    By redevo7 in forum Statistics
    Replies: 1
    Last Post: 04-05-2006, 10:43 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats