The Kruskal Wallis Test By: JOSEPH PINES •What is Kruskal-Wallis Test? The Kruskal-Wallis test is a nonparametric statistical method used to determine if there are significant differences between three or more independent groups WHY DO WE USE KRUSKAL WALLIS TEST? The data is not normally distributed. The sample sizes are small. We want to compare medians between groups. . • 12 H= 𝑁 𝑁+1 𝑘 𝑖=1 𝑅𝑖 2 − 3 𝑁+1 𝑛𝑖 Formula and Symbols: •H = Kruskal-Wallis statistic •k = number of groups •N = total number of observations •Ri = sum of ranks for group i •Ni = number of observations in group i . • Problem • You want to compare the effects of three different study methods on student exam scores. The scores for each method are as follows: • Method A: 78, 85, 88 . • Method B: 65, 70, 72, 75 • Method C: 80, 82, 84 •Step 1: Rank all the data. • Combined SCORES RANK data:78,85,88,65,70,72,75,80,82,84 65 1 70 2 72 3 75 4 78 5 80 6 82 7 84 8 85 9 88 10 •Step 2:Calculate Ranks and Group Statistics Ra=5+9+10= 24 Na = 3 Rb=1+2+3+4= 10 Nb = 4 Rc=6+7+8= 21 Nc = 3 • Step 3: Calculate the Test Statistics (H) 𝑘 12 H= 𝑁 𝑁+1 𝑖=1 𝑅𝑖 2 − 3 𝑁+1 𝑛𝑖 . Where: N= Na + Nb+ Nc = 3 + 4 + 3 = 10 K= 3 (number of groups) • • Step 4: Calculate the Kruskal-Wallis statistic 𝑘 H. 𝑅𝑖 2 − 3 𝑁+1 𝑛𝑖 12 𝑖=1 H= 𝑁 𝑁+1 𝑘 242 102 212 ( + + ) − 3 10 + 1 3 4 3 12 𝑖=1 H= 10 10+1 . 𝑅𝑖 2 1. Calculate: : 𝑁𝑖 242 576 Method A: = = 192 3 3 102 100 Method B: = = 𝟐𝟓 4 4 212 441 • Sum of these values: 𝑘 𝑖=1 𝑅𝑖 2 = 192 + 25 + 147 = 364 𝑛𝑖 2. Substitute values into H: 12 H= (364) – 3 (10+1) 10 (11) 12 H= (364) – 3 (11) 110 4368 H= – 3 (11) 110 4368 H= - 33 110 H= 39.71 – 33 = 6.71 •Degrees of Freedom: df = k- 1 df= 3 – 1 = 2 For df = 2 and alpha (α) = 0.05 the critical value from the Chi-squared distribution table is approximately 5.991. Conclusion: Calculated H : 6.71 Critical Value: 5.991 Since H: 6.71 > 5.991, we reject the null hypothesis There is significant evidence to suggest that there are differences in exam scores among the three study methods. •Activity Problem: • A researcher is studying the effect of three different diets on cholesterol reduction after a 10-week program. The cholesterol reductions (in mg/dL) for each diet group are as follows: • Diet A: 12, 15, 18 • Diet B: 10, 11, 14, 16 • Diet C: 8, 9, 13 The end! Thank you!