Project description

This project ties together what you learned about:

  • the importance of exploring data graphically
  • the benefits of randomization in conducting experiments
  • choosing the correct analysis for a given situation
  • interpreting the results of hypothesis tests and confidence intervals when comparing two groups
  • writing up the results of a statistical analysis

Your group will be graded on how well it performs with respect to each of the above points. I will be more than happy to answer broad questions about these statistical points and how to do something in Minitab, but not direct questions about how to complete any part of this project.

The project

A randomized experiment was conducted to see whether the information that a person is provided affects his/her estimate of the population of a country.  Each student in the experiment estimated the population of the countries Turkey and Canada.  However, the information provided to the students on the data collection form differed.  A random half of the students received a green form, which contained:

  1. Do you think the population of Turkey is more than 80 million?   Yes _____ No _____
  2. To the nearest million, estimate the population of Turkey:  ______ million.
  3. The population of Australia is about 18 million.  To the nearest million, estimate the population of Canada:  _____ million.

and the other half of the students received a blue form, which contained:

  1. Do you think the population of Turkey is more than 10 million?   Yes _____ No _____
  2. To the nearest million, estimate the population of Turkey:  ______ million.
  3. The population of Australia is about 18 million.  To the nearest million, estimate the population of Canada:  ____ million.

That is, before estimating the population of Turkey, half of the students saw the number "10 million" and half of the students saw the number "80 million."  And, before estimating the population of Canada, all of the students saw the same number, namely "18 million."

The data from this experiment are stored in: ljs_03a.txt. Because the experimenter was concerned that some confounding factors could cloud the study results, the experimenter collected additional information on each student, including the student's GPA, whether or not the student had ever been to Canada, and whether or not the student had ever taken geography. The data can be described as follows:
 
Column name Description
Form 1 = Green, 2 = Blue
Turkey Student's guess of population of Turkey
Canada Student's guess of population of Canada
GPA Student's GPA
Geography Have you ever taken a geography class? 1 = Yes, 0 = No
InCanada Have you ever been to Canada? 1 = Yes, 0 = No

Analyze the data to determine whether survey responses can be affected by extraneous information on the data collection form. Write a conclusion of this experiment and your statistical analysis. Include any relevant Minitab output. In your conclusion, make sure you write the strongest conclusion you possibly can. Specifically address (but not necessarily in this order):

  • What was trying to be learned by conducting the experiment, and how the experiment was conducted?
  • What you learned by graphing the data?
  • Besides the difference in the forms they received, do the groups differ in any other way? And, if so, how? (It would be useful to summarize the differences -- or similarities -- in a table.) If the groups don't differ except for the forms they received, comment why.
  • If you can say the guesses differ between the two forms, indicate by how much.
  • If you can say the difference in the forms caused the differences in guesses, comment why.

Please submit a double-spaced report.

Useful resources

What to hand in?

Submit a typewritten, double-spaced report. On the top of your report, make sure you include your group number, and each group member that participated in the project.

Deadlines

No exceptions -- all projects must be turned in at the end of the lab.