Hi all
I want to compare different data sets but cannot access raw data. Only the averages and standard deviations are available. How can I do tests such as Duncan or Tukey?
For example:
Group A: 25±2
Group B: 35±4
Group C: 28±3
Group D: 41±2
To perform post anova test like Duncan or Tukey, it’s necesary to have obviously, the anova result, because for their performance you need to know the estimation of the mean square for the experimental error (including the number of subjects for each group).
Only with the mean values and standard deviation, you can’t do any statistical hypothesis test comparation, only its possible a descriptive comparison.
Best regards.
Dear Amir if you have the volume number of your set it is o.k.
You have the standard error and you can make t test between the paired samples
Making repeated or consecutive t tests with the same set of groups (A-B, A-C. A-D. B-C, etc) increases the probability of type I error, and its not recommendable. The post-anova tests prevents this problem, for what I insist that anova results are necessary to perform this comparisons.
For other hand, the standard errors will be many as number of comparisons (for pairs) that will perform with t test. The anova estimates only one standard error that affects to all groups (mean squared error), that its usefull for the post-anova comparisons.
Best regards.
- Badges
- Science method