Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
t-test vs paired t-test
Message posted by Andrea Laliberte (via 65.0.120.30) on August 23, 2001 at 4:17 PM (ET)
I am testing a technique to count objects using a computer. In order to assess how well this works, I am comparing the computer counts from the image with manual counts from the same image. I have 10 images. Is it appropriate to use a paired t-test or a 2 sample t-test?
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: t-test vs paired t-test
Message posted by darius (via 200.23.217.10) on August 23, 2001 at 4:53 PM (ET)
Use a paired t-test when the same thing is tested by two methods (sample 1 of method 1 is the same as sample 1 of method 2).
Re: t-test vs paired t-test
Message posted by Phil (via 216.175.115.253) on August 25, 2001 at 1:48 AM (ET)
The t-test for independent events is used when you are using two different samples. If you are measuring the same sample, but using two different methods,conditions, or times, then used the paired t-test...they are not independent.Ho: difference=0
Ha: difference not=0
So, I think you should use the paired t-test.
Re: t-test vs paired t-test
Message posted by Keith M. Bower (via 12.4.141.8) on October 1, 2001 at 4:35 PM (ET)
I agree with Phil. Where a dependency exists (e.g. left shoe and right shoe for each child in a study) it is appropriate to perform a paired t-test. I've got a paper on it at http://www.minitab.com/company/virtualpressroom/Articles/index.htm if you're interested.
t-test vs paired t-test
Message posted by Dionisis (via 155.245.254.253) on October 21, 2001 at 4:47 PM (ET)
We have 20 randomly selected people.ten of them took a pill A and the other 10 took a pill B.We don't know who person took which pill.We have the numbers of their blood presure in the beginning of the experiment and the numbers at the end and the numbers of the decrease of the blood presure. We carry out the t-test in the beginning.The the problem says that we are informed that the data were actually paired.The men supposed to paired according to their original blood pressure measurements i.e the man with low initial pressure should have been paired with another man who have low initial pressure. We want to know now whict t-test is appropriate!
Could you give me your opinion please?
thanks!
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.
