Projects in Statistics Courses

Mary Parker, prepared for MAA Minicourse, January 18, 2000.

http://www.austincc.edu/mparker/stat/maajan2000.html

 

Goals for projects:

To give students experiences in:

  1. Thinking of how they could use statistics in their lives and professions.
  2. Defining their question and defining appropriate variables to measure.
  3. Deciding upon a sampling scheme or an experimental design.
  4. Producing data.
  5. Finding data from an original source rather than data that is already prepared for use in a statistics class.
  6. Preparing a report appropriate to give to a manager to support decision-making.
  7. Recognizing that even if they have done the project well, there will still be more that could be done in gathering and analyzing data to get information relevant to the question.

Issues that are difficult for students:

  1. Thinking of a question for which the answer might make a difference in a decision.
  2. Recognizing that ambiguities in defining or measuring the variable are difficulties to overcome, not reasons to quit working on the problem.
  3. In sampling, dealing with the trade-off between wanting a substantial or important population and needing to do a simple random sample.
  4. In sampling, determining how important it is to take a simple random sample, as opposed to a sampling scheme that seems as if it could be thought of as a simple random sample. (Eg. In sports statistics, we might predict the entire season average from the average in the first few games. Is that a simple random sample? Is it generally accepted as appropriate statistical inference?)
  5. Finding the time to do data collection individually or in their group.
  6. Dealing with the ambiguities in defining or measuring the variable that often arise in the process of data collection.
  7. Writing a report in a form suitable for professional work (as opposed to the usual homework style of presenting a solution.)
  8. Recognizing that even if they have done the project well, there will still be more that could be done in gathering and analyzing data relevant to the question.

Dealing with the difficulties:

  1. Require that students work in groups so that they will discuss the difficulties together and maybe resolve some of them.
  2. Make examples of previous successful projects available to students.
  3. Have students critique others' plans for their study. (Provide rather specific guidelines, of course.)
  4. For a big project, require a first draft and give feedback. (Maybe even multiple drafts with feedback.)
  5. Require each group to meet with the instructor to discuss the project.
  6. Split a larger project up into smaller projects.
  7. After the students have finished a big project, have them include a section about "future study," to emphasize that suggesting areas for further work is part of a good study, not an admission that they did something wrong.


Websites with material relating to projects:

Journal of Statistics Education: Articles on Projects (Browse the index. Fillebrown, Mackisak, and Smith.have articles relevant to projects in general introductory courses.) http://www.amstat.org/publications/jse/jindex.html

Beth Chance's syllabus (with project assignments) in her elementary course this semester. http://statweb.calpoly.edu/chance/stat217/syllabus.html

Beth Chance and Anne Sevin's Peer Review Worksheet

Mary Parker's project assignments in her elementary course this semester. http://www.austincc.edu/mparker/1342/projects.htm

Other material relating to projects:

Moore, Thomas, et. al. (2000) Teaching Resources for Undergraduate Statistics, to be published by the MAA Notes series in 2000. (Includes lots of material, including some on projects.)

Ledolter, J. (1995) "Projects in Introductory Statistics Courses" The American Statistician, Volume 49, Number 4, p. 364-367

Roberts, H. (1992) "Student-Conducted Projects in Introductory Statistics Courses", in Statistics for the Twenty-first Century, MAA Notes, no. 26.


Examples of short projects to be done early in the semester:

Project 1:

Recall the example about cats that we discussed in class. My question was whether my cat is particularly big. The cases (also called individuals) were cats, and the variable we chose was "length", in inches, measured to the nearest inch. Recall that we had to define the variable more precisely. Some questions were (1) Should I include the tail? and (2) If I include the tail, do I measure to the tip of the fur or just to the tip of the tail without the fur? Observations on the length of a cat might be 15 inches, 18 inches, 12 inches, and 13 inches.

Now, you do one. Think of some variables which you might be interested in analyzing sometime.

  1. What is the question you’re interested in investigating?
  2. Describe what a single case is.
  3. Describe three different variables that could be measured on a case.
  4. Pick one of the three variables (underline it) and describe how you would measure it and what the units will be. Are there likely to be any problems or ambiguous situations?
  5. For that variable, give four observations that would be reasonable to see.


Project 2: Plan a study.

1. Formulating the question. Think of a question of interest to you for which you (you, as a student and an individual) could collect some data.

Examples:

a. What percentage of cars don't completely stop at the stop sign at the north corner of Waterston and Hartford?

b. In a blind tast test, do people prefer Classic Coke to Pepsi?

c. How accurately could we predict hat size from forearm length?

Write the question and explain how this question is of interest to you or someone. How might accurate knowledge about the answer change someone's actions?

2. Identifying relevant variable(s).

Explain what the variable(s) is and how you expect to measure it. Discuss any definitions needed. (Remember the example about the length of a cat -- does that include the tail or not?)

3. How would you collect useful data?

Is this a question for which you just go collect data or do you have to set up an experiment? (Example b above requires you to set up an experiment.) Explain how you could get an appropriate sample or design an experiment. Describe the population in which you are interested and explain how this method of collecting data will result in data from this population that is appropriate for statistical inference.

 


Last updated January 17, 2000. Mary Parker, mparker@austincc.edu