Chapter 7: Non-Parametric Tests

Question 1

what are non parametric tests?

Answer 1

• used when we’re not sure that the sampling distribution is normal
• “assumption free” tests
• have the same assumptions as their parametric counterparts, but if you violate those assumptions, you can still trust the test

for tests of normality/homogeneity, you do not want to find sig results. sig results (p <.05) indicate violation of assumption

Question 2

mann whitney test

alternative to independent samples t test

Answer 2

• used in situations when you have 2 groups and you have different entities in those 2 groups and want to compare those groups on 1 quantitative variable

r = z/ sq rt N

always check to see if you can do parametric version first. these are preferred if assumptions are met since parametric tests are more powerful than nonparametric

Question 3

mann whitney results in APA

Answer 3

Depression levels in ecstasy users (Mdn = 17.50) did not differ significantly from alcohol users (Mdn = 16.00) the day after the drugs were taken, U = 35.50,z = -1.11, p = 0.280, r = -0.25. However, by Wednesday, ecstasy users (Mdn = 33.50) were significantly more depressed than alcohol users (Mdn = 7.50), U= 4.00, z = -3.48,p < .001, r = -0.78.

U statistic

Question 4

wilcoxon signed rank test

alternative to dependent samples t test

Answer 4

• used when you want to compare 2 groups with the same entities on a quantitative variable. you are measuring the same people over time

r = z/ sq rt N

remember to double N since each person tested twice (i.e., if you have 10 people, then N is 20)

Question 5

wilcoxon signed rank test in APA

Answer 5

For ecstasy users, depression levels were significantly higher on Wednesday (Mdn = 33.50) than on Sunday (Mdn = 17.50), T = 36, p = .012, r = 0.63. However, for alcohol users the opposite was true: depression levels were significantly lower on Wednesday (Mdn = 7.50) than on Sunday (Mdn = 16.0), T = 8, p = .047, r = -0.44.

T statistic

Question 6

kruskal wallis test

alternative to one way independent ANOVA

Answer 6

• people in different groups, and you want to compare them on some quantitative variable
• can be expanded to 2 or more groups
• if test is not significant, then stop.
• if test is significant, do a pairwise comparison or step-down procedure to determine which groups are different. this helps keep FW alpha at .05 since you do not have to adjust p values as much, which gives you more power to find differences that might exist between dif groups

r = z/ sq rt N

Question 7

kruskal wallis in APA

Answer 7

energy levels were significantly affected by eating fast food, H(3) = 8.66, p = .034. Pairwise comparisons with adjusted p-values showed that there were no significant differences between energy levels when people ate 7 fast food meals per week compared to 4 meals (p = .133, r = 0.36), 1 meal (p = .133, r = 0.36) or no meals (p = .058, r = 0.41). There were also no significant differences in energy levels between those eating 4 fast food meals per week and those eating 1 meal (p = 1.00, r = 0.00) and no meals (p = 1.00, r = 0.05). FInally, there were no significant differences in energy levels between those eating 1 fast food meal per week and those eating none (p = 1.00, r = .05).

H statistic w df

Question 8

friedman’s ANOVA

alternative to repeated measures ANOVA

Answer 8

• compare multiple groups on one quantitative variable and you have the same people
• like signed rank, but you can have more than 2 groups
• if test is not significant, then stop. (fail to reject the null, the null is tenable)
• if test is significant, do a pairwise comparison or step-down procedure to determine which groups are different. this helps keep FW alpha at .05 since you do not have to adjust p values as much, which gives you more power to find differences that might exist between dif groups

r = z/ sq rt N

Question 9

friedman’s ANOVA in APA

Answer 9

• the weight of participants did not significantly change over the two months of the diet, χ2(2) = 0.20, p = 0.91.