Agriculture Reference
In-Depth Information
friedman block*
Friedman = 11.0714
Kendall = 0.5536
p-value = 0.0500
Notice how the variable ( block* ) is entered. Because each vari-
able has the same root name ( block ), adding the asterisk indicates
to the command to use all variables with this root. The command
could have been entered as friedman block1 block2 block3
block4 , and would have resulted in the same outcome. The chi 2
value is labeled as Friedman, which is 11.0714. The p-value of 0.0500
indicates there are differences between treatment medians at the 5%
level. The ANOVA for these data was highly significant with an F
value of 4.83 and a p-value of 0.0080, again showing the greater
power of parametric tests when the data do not grossly violate the
assumptions. The Kendall number is a value between 0 and 1. The
Kendall W, or coefficient of concordance, is an indicator of how well
the blocks agree in their ranking of the treatments with higher values
indicating greater agreement.
As mentioned throughout this chapter, nonparametric statistics
are generally not as powerful as parametric statistics. They are useful,
however, where the underlying samples don't meet the requirements
of parametric tests and a reasonable transformation is not available.
Search WWH ::




Custom Search