Statistics Chapter 15 (Non-parametic Statistics)

Question 1

What is the basic idea of non parametic testing

Answer 1

By ranking the scores with each other you can get the sampling distribution of the test statistics without an assumption about the normal probability distribution of the response variable

Question 2

How does the Wilcoxon test work

Answer 2

The Null hypthesis is that the population distributions are the same (average rank is the same)
You then take the scores and rank them
The average rank in each condition may be different
You get the P value by comparing the test statistic with all the possible rankings that COULD have been

Question 3

How does the Kruskal-Wallis test work

Answer 3

You find the average rank in each condition and compare them to the average overall rank with the formula
This gives you the p-value for the X^2 table

Question 4

How does the Sign test work

Answer 4

You compare the two matched groups or condition and take the sign of the difference (+/-)
The Nullhyothesis is that that the proportions of + is 0.5 (independent pairs)
Test statistic is number of subjects with +
The P-Value us the sum of probabilites of observing a higher test statistic

Question 5

When do you use non-parametic analysis

Answer 5

1. When the measure is ordinal
2. We dont know if data is normally distributed
3. Data is not normally distributed and sample size too low for CLT