Hello,
If we are comparing lethality between our wild type and mutant or doing locomotion assays (thrashes; body bends) between them, which one we should use -
Unpaired OR paired two-sample t-tests?
Thanks.
Hi,
unpaired t tests are where you have two independent groups (let’s say control and test) and you test whether the means of each sample differ enough statistically to suggest they represent two different populations. both cases assume a normal distribution and that the sd is more or less the same for each group.
In your example, wildtype and mutant are the two independent groups and thrashing rate or lethality could be the test.
In a paired t test you would test one group twice and compare the means to see if they differ significantly.
In your example, this would be slightly different to the unpaired t test because essentially you would be assessing if an intervention (drug/stimulus) in your wildtype OR your mutant strain changed a variable such as thrashing rate…
Hope this helps?
Steve
Here are some great resources for stats, what they can and cannot do, and when to use which tests:
http://www.graphpad.com/support/faqid/1790/
http://www.graphpad.com/guides/prism/6/statistics/
also useful reading:
http://jcb.rupress.org/content/177/1/7.full
Wow…
Thanks a lot everyone…
