Power point on non parametric tests
-
Upload
kevin-hendry -
Category
Documents
-
view
271 -
download
4
Transcript of Power point on non parametric tests
FIRST, REMEMBERParametric tests has greater power to detect a
genuine effect if one exists than a non-parametric test.
But this is true only if assumptions are met.
EXPLORE YOUR DATA FOR BIASES
Analyze
Descriptive Statistics
Explore
If p < .05 for any group ( sign. different distribution of scores), then it’s recommended to conduct a non-parametric test for all groups being compared.
GENERAL PR OC EDURE S OF NON-PARAM ET RIC TE ST S
Analyze
Non-parametric tests
Independent Samples or Related Samples
WHAT TEST DO I USE?For a basic independent t-test:
• Mann-Whitney test or Wilcoxon’s Rank-sum test
For a basic paired samples t-test:• Wilcoxon signed-rank test
• also called Wilcoxon matched-pair signed ranked
For several independent groups (ANOVA):• Kruskal-Wallis test (rank-sum design)
For several dependent (related) groups (ANOVA):• Friedman’s ANOVA (signed-rank design)
THEORY OF RANK-SUMAll non-parametric tests overcome the shape of the distribution of scores by ranking the scores in your data. Score 3 5 6 6 7 8 9 10 17 24 27 28 29 30 32
35 35 35 36 39 Potential Rank
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Actual Rank
1 2 3.5 3.5 5 6 7 7 9 10 11 12 13 14 15 17 17 17 19 20
Group A A A A A A A A A E E E E A E E E E E E
Sum of ranks for alcohol (A) = 59 sum of ranks for ecstasy (E) = 151
THEORY FOR SIGNED-RANK Star
tMonth 1
Month 2
Start(Rank
s)
Month 1
Month 2(Ranks)
Person 1
63.75
65.38 81.34 1 2 3
Person 2
62.98
66.24 69.31 1 2 3
Person 3
65.98
67.70 77.89 1 2 3
Person 4
107.27
102.72
91.33 3 2 1
Person 5
66.58
69.45 72.87 1 2 3
Person 6
120.46
119.96
114.26 3 2 1
Person 7
62.01
66.09 68.01 1 2 3
Person 8
71.87
73.62 55.43 2 3 1
Person 9
83.01
75.81 71.63 3 2 1
Person 10
76.62
67.66 68.60 3 1 2
Ri 19 20 21
DEMONSTRATE TEST OFTWO RELATED CONDITIONS
The non-parametric equivalent of the paired-samples t-test
Using quail sex study: Matthews et al. (2007).sav
Switch to SPSS and continue…..
WHAT IS THIS STUFF?
z-scorep-value
DESCRIBING YOUR CONCLUSIONIn common terms, conditioning (as a learning mechanism) provides some adaptive benefit in that it makes it more likely that you will pass on your genes.
In more scientific terms; of the 78 eggs laid by the test females, 39 eggs were fertilized. Genetic analysis indicated that 28 of these (72%) were fertilized by the signalled males, and 11 were fertilized by the control males. Ten of the 14 females in the experiment produced more eggs fertilized by the signalled male than by the control male, T = 13.5, p < .05). These effects were independent of the order in which the 2 males copulated with the female.
In addition, the present findings show that when 2 males copulated with the same female in succession, the male that received a Pavlovian CS signalling copulatory opportunity fertilized more of the female’s eggs. Thus, Pavlovian conditioning increased reproductive fitness in the context of sperm competition
WHAT AB O UT EFFECT S IZ E?
r = z-score ÷ √ N
r = -2.23 ÷ √ 14
r = -2.23 ÷ 3.74 = .59
.59 is a large effect.
SEVERAL IN DEPEND EN T GRO UPS
The Kruskal Wallis test
Based on ranked data, like the Mann-Whitney test
Rank scores from lowest to highest, ignoring the to which the score belongs, then assign the lowest a rank of 1.
DATA TABLE EXAMPLENo Soya 1 Soya
Meal4 Soya Meals
7 Soya Meals
Sperm Rank
Sperm
Rank
Sperm
Rank Sperm
Rank
0.35 4 0.33 3 0.40 6 0.31 10.58 9 0.36 5 0.60 10 0.32 20.88 17 0.63 11 0.96 19 0.56 70.92 18 0.64 12 1.20 21 0.57 81.22 22 0.77 14 1.31 24 0.71 131.51 30 1.53 32 1.35 27 0.81 15~~~~ ~~
~~~~ ~~
~~~~~
~~~~
~~~ ~~~
Total Ri
Average
92746.35
88344.15
88344.15
54727.35
CON DUCTING AN EXPL OR ATO RY AN ALYSIS
The Kruskal-Wallis test tells us that, overall, groups come from different populations, but not specifically where they differ between the different groups.
You could do pairwise comparisons, testing each group separately, but it would inflate the Type I error at .05.
The alternative is to use a stepped procedure• By ordering the groups based on the sum from lowest to highest.
The data editor has two columns, one coding variable & one dependent variable
File Soya.sav,
EXAMPL E OF ENTRY FOR EXPL O RING DATA
Dependent variable
Coding variable
Q - Q PL O TS, CLEAR DEVIATIO N IN AL L GRO UPS
Each group has less than n = 30, all have non normal distribution, so use non-parametric test
SPSS PROCEDURES
Analyze
Non-parametric testsIndependent Samples
DON’T FORGET EFFECT SIZE
r = z-score ÷ √ N
r = -2.476 ÷√ 80
r = -2.476 ÷ 8.94
r = -0.28 small to medium reverse effect size
ANY QUESTIONS?
Fewer questions are positively correlated to quicker end of class time.