When should researchers use different hypothesis tests for independent and related samples? Is one type of sample preferable over the other? Under what circumstances. Give an example of a problem that represents independent samples and one that would represent dependent samples.