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

1. ## 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. 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. 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)?

4. 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. 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. Originally Posted by Gina Gina Gina
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

 Tweet

#### Posting Permissions

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