What role does the binomial distribution play in the sign test, ...

[Bob-Dole]

What role does the binomial distribution play in the sign test, and when is it permissible to use the normal approximation in carrying out the test?

Question 2

Compare the Friedman test with the randomized block ANOVA in terms of: (a) their respective null and alternative hypotheses, and (b) the assumptions required in order to use each test.

[wuly]

Answer to Question 1

The sign test relies on the fact that, if the true population (of differences) median is equal to zero, there is a 0.50 probability of obtaining a positive difference. Thus, this is a binomial experiment and a difference will be either positive or negative (ties are ignored). When the number of observations (non-zero differences) is at least ten, the normal approximation may be used.

Answer to Question 2

a) The null and alternative hypotheses for the Friedman test and the randomized block ANOVA are almost identical except that the Friedman test deals with medians and the ANOVA test deals with means. b) Randomized block ANOVA requires that observations be from normal populations with equal variances and that the observations be measured on interval or ratio scale. The Friedman test makes no assumptions about the underlying populations, and the observations may be measured on the ordinal, interval, or ratio scale.

[Bob-Dole]

Correct answers!

[wuly]

Great! Please up vote