Group Project Guidelines for Statistics 528

For this project you should write a brief (about 3 or 4 pages
including figures and parts of any computer output) report outlining
the results of a statistical study that you and your teammate(s)
will conduct.  Your statistical study should include the following
steps:

  1. Formulate a question that can be answered by collecting data in an
     experiment, observational study, or other appropriate method.

  2. Decide what data will be collected (i.e. what variables are to be
     controlled, measured or observed).

  3. Decide how to collect the data (observational study,
	experiment, etc.).

  4. Decide how much data are to be collected (i.e., the sample size) and
	explain why (e.g., not enough time or money for a larger sample).

  5. Collect the data or conduct the experiment.

  6. Analyze the data. This should include graphical analysis and other
     analyses, as appropriate.

  7. State any assumptions that you need to make. If these assumptions can
     be checked, describe how you did this. If they cannot be checked
     explain why you feel comfortable making the assumptions.

  8. State clearly the conclusions that you can draw from your analysis.

The written part of the project should be done on a word processing
system.  Students must work in their groups to complete the
project. One grade will be assigned for all students in each group.



Here is a summary of the form that your project data should have:

There should be a response (Y)  variable that is numeric.

There should also be one or more explanatory (X) variables. These
variables may be either numeric or nominal (categorical). The X
variables should have some reasonable chance to explain some of the
variability on the Y variable.

Your study could be either observational or experimental. If your
study is observational, you should explain why it was not feasible
to conduct an experiment.

The number of observations is not critical, but it will be easier to
identify a model if you have at least 15 or 20 observations
(depending on the amount of spread in the data). Generally more
observations is better, but working with really large data sets is
something that we have not done in course work, and analysis of data
sets with thousands of observations may be inconvenient.

It is always better to work on an example that is of some interest
to the analysts. Data may be available from the Internet.

If you use data that someone else has collected, then you must
provide the exact source of the data, provide a copy of that source,
and perform some analyses that go beyond what is in the original
source.

Here is a list of the Project milestones. See the Important dates
file for the due dates for the milestones.

As a general guideline for this project, let me suggest that some of
you might want to conduct a "pilot" study to try out your data
collection technique(s). In practice it is common to spend about 10
to 15% of ones resources on such a pilot study in an effort to
gather preliminary information so that the "main study" will have a
better chance of being successful. Often adjustments are made to the
study between the pilot and the main study (especially, for example,
changing the wording of questionnaires of deciding to take data on
other variables that may be observable).


Project milestones (include the names of all team members
on each mile stone):

Milestone 1: Initial project proposal          

    Brief description of the purpose of the project and the planned
    data source. 



Milestone 2: Detailed project proposal      
    
    A statement of the question that you hope to answer
    with your study. Explain more detailed description
    of how you will gather the data to complete the study.
    Will you use an experiment or an observational study?
    What are the "observational units" in your study?
    Which variable(s) will you record for each observational units?
    How have you decided how much data to collect (i.e., how
    did you choose how many observational units to choose?)?



Milestone 3: Data Summary (hard copy of tables and/or graphs) 

    Provide a brief summary of the data that you have collected
    for either the main study or your pilot study (if you did one).
    The summary should consist of either summary tables  or
    graphs (which ever is more appropriate) and initial conclusions. 
    If you are reporting
    the results of the pilot study, indicate how the results of
    the pilot study are affecting the main study. The graphs
    and tables at this stage should be neat, but they do not have to
    be done on the computer.


Final project due