Question: One experiment contrasted responses to "prediction...
One experiment contrasted responses to "prediction-request" and to "request-only" treatments, in order to answer two research questions.
i. Can people predict how well they will behave?
ii. Do their predictions influence their behavior?
In the prediction-request group, subjects were first asked to predict whether they would agree to do some volunteer work. Then they were requested to do the work. In the request-only group, the subjects were requested to do the work; they were not asked to make predictions beforehand. In parts (a-b-c), a two-sample z-test may or many not be legitimate. Ir it is legitimate, make it. If not, why not?
a. 46 residents of Bloomington, Indiana were chosen at random for the "prediction-request" treatment. They were called and asked to predict "whether they would agree to spend 3 hours collecting for the American Cancer Society if contacted over the telephone with such a request." 22 out of the 46 said that they would. Another 46 residents of that town were chosen at random for the "request only" treatment. They were requested to spend the 3 hours collecting for the American Cancer Society. Only 2 our of 46 agreed to do it. Can the difference between 22/46 and 2/46 be due to chance? What do the data say about the research questions (i) and (ii)?
b. Three days later, the prediction-request group was called again, and requested to spend 3 hours collecting for the American Cancer Society: 14 out of 46 agreed to do so. Can this difference between 14/46 and 2/46 be due to chance? What do the data say about the research questions (i) and (ii)?
c. Can the difference between 22/46 and 14/46 be due to chance? What do the data say about the research questions (i) and (ii)?