Chapter14 Nonparametric Statistics
Parametric statistical tests: methods for making inferences about population parameters
(confidence interval and hypothesis testing) rely on the assumptions about probability
distribution of sampled populat
14.6 Comparing three or more populations: randomized Block design
Friedman Fr -test for RBD: no assumptions concerning the population probability
1. Rank observations within each block;
2. Add up the ranks for each treatment
14.7 Rank Correlation
In chapter 11, we have introduced a correlation coefficient r based on the observations of
two numerical variables.
Here we introduce a correlation coefficient based on ranks of observations:
Spearmans rank correlation coefficient rs
Distribution-free tests statistical tests
that dont rely on assumptions about the
probability distribution of the sampled
Nonparametrics branch of inferential
14.5 comparing three or more populations: completely randomized design
Nonparametric method can used to analyze the data from CRD.
Kruskal-Wallis H-test for CRD: no assumptions concerning the
population probability distribution.