Question 464762
1. A parametric test is based directly on statistics derived from a distribution with parameters, such as the normal distribution.  A nonparametric test is not based on normal distribution and is often based on the rules of probability and counting.  They are usually slightly less statistical power than the equivalent parametric test.

2. The Kruskal-Wallis Test is a form of ANOVA that is a non-parametric test that does not assume nomality. It compares multiple groups based on ranking. The ANOVA does base results on assumption comparing multiple groups based on rank.


Step 1)


H0 = The feed is from the same population of turkeys.
H1 = The feed is from different populations of turkeys.

Step 2)

alpha = .05

Step 3)

Degrees of Freedom = 2

Step 4) Decision
Reject the null hypothesis when the critical value is greater than 6.98.

Step 5)
The null hypothesis is rejected as the value of all feeds are greater than 6.98.  Therefore all feeds are provided by different populations.
Kruskal-Wallis Test 	  	  	 
  	  	  	  	  	 
  	Median 	n 	Avg. Rank 	  	 
  	11.30 	5 	8.80 	 Feed mixture A
  	11.15 	6 	7.67 	 Feed mixture B
  	11.90 	5 	9.20 	 Feed mixture C
  	11.30 	16 	  	Total 	 
  	  	  	  	  	 
  	  	  	0.315 	 H (corrected for ties)
  	  	  	2 	 d.f. 	 
  	  	  	.8541 	 p-value 	 
  	  	  	  	  	 
  	  	multiple comparison values for avg. ranks
  	  	6.98 	(.05) 	8.56 	(.01)