Your data analysis solution
Name is required.
Email address is required.
Invalid email address
Answer is required.
Exceeding max length of 5KB

Pairwise comparisons

By Natalia | Jan 12, 2018 09:08PM CET


Doing an ANOVA test and a Fisher LSD pairwise comparison on a set of 10 variables in 3 observations (3 reps each), I have realized that the output of the Fisher test is different if I run the 10 variables at once that if I run them one by one. The parameters selected for the analysis are the same cause I am just changing the selection of the dependent variables. Does it mean that I have to do the variables one by one?
Another question is... In the output of the Fisher test, it shows a mean and a standard error; however, this "mean" is not a real average of the 3 reps of each observation and the standard error is the same in the 3 observations (which it's not true). What does this "mean" really mean?





By Efthalia | Jan 17, 2018 11:42AM CET | XLSTAT Agent

Hello Natalia!

Fisher LDS results should be the same whether you run the 10 ANOVAs separately or all at once. It is quite surprising to get different results. Could you please send me your data and XLSTAT output at Then I will have a look at the issue.

Regarding your second question, if the mean value does not correspond to the average of the observations then you are probably using the least squares means. Check out “here” the difference between LS Means and Observed Means.


Post Your Public Answer

Contact Us
seconds ago
a minute ago
minutes ago
an hour ago
hours ago
a day ago
days ago
Invalid characters found