0
$\begingroup$

Regarding my data collection, I have five groups plus a control group. I need to compare each group with control condition with and then find the highest difference in one of these groups.To briefly say, (Group1-Control group), (Group 2-Control Group) ...Group 5-Control Group).

May I use Paired sample T-Test and then find the significant level of paired groups?

$\endgroup$
  • 1
    $\begingroup$ No, because the results will be correlated (due to their reliance on a common control group). Search our site for ANOVA. $\endgroup$ – whuber Nov 29 '16 at 23:19
  • 2
    $\begingroup$ You may be confusing the concepts of paired sample (each group consists of the same number of identical individuals measured under another condition) and pairwise tests (you have multiple groups of some sort and compare them in pairs) $\endgroup$ – David Ernst Nov 29 '16 at 23:48
0
$\begingroup$

If you are thinking about T test, you have a continuous outcome variable. If you need to compare each control group with a reference /control group you can do it by means of Dunnett’s Multiple Comparison. This can be ran after ANOVA in case of overall significance. However, this will only tell you what experimental groups are significantly different from control. If you want draw conclusion on the superiority of a particular group you should run pairwise (unpaired) T test.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.