Langston, Research Methods Laboratory, Exercise 2 -- Measurement/Ethics

Part 1:  Statistics

For each research question, formulate a hypothesis, predict an outcome, and tell me what statistical procedure would be appropriate to test your hypothesis:

1.  Will a good night's sleep help comprehension in an early morning class more than staying up and watching TV?  I have some people sleep 8 hours one night, and then the same people watch TV the next night.  I measure performance in the morning class after both treatments.

2.  Will "cramming" all night lead to better performance on exams than sleeping for a few hours?  I have one group study from 8 PM until midnight, and a different group studies from 8 PM all night.

3.  How much alcohol does it take to impair driving?  I have one group take one drink, a different group take two drinks, and a different group take three drinks.  Then I have each group drive an obstacle course.

4.  Will hot weather increase violent behavior?  I measure temperature and number of violent acts for each day of the summer and look at the relationship between temperature and violence.

Part 2:  Ethics

Discuss the ethical implications of these experiments:

1.  What happens to the results of an experiment when the experimenter has an expectation in mind?  We get some students and tell them that we're replicating recent work on the relationship between race and IQ.  Then we give each of them a hypothesis to work with.  The thing is, none of these people are experimenters.  The real test is to see if the data they collect confirms their hypothesis.  (The design is similar to Rosenthal, 1966.)

2.  What happens when a "normal" person is placed in a mental institution?  I have some "patients" complain of existential psychosis (they report hearing voices saying "empty," "hollow," and "thud").  This is the only symptom, otherwise they tell the truth and respond faithfully on all tests.  After the first day, they show no additional symptoms.  Will the staff of the institutions recognize that these people are not mentally ill?  (The design is similar to Rosenhan, 1973.)

Research Methods Lab Exercise 2
Will Langston