This project gives you the opportunity to employ your skills in statistical analysis to address a question that you find interesting. You will collect the data, perform the statistical analysis, and present your conclusions to the class. You will work with two partners; you should contribute equally to the project, and will receive the same grade.
The project will consist of 5 steps:
1. Submit a question you plan to address (due
second Monday).
This should be brief - two or three sentences. Examples of
questions
you might ask are:
Do CC students share
the same political beliefs as their parents?
How does tree trunk diameter vary with distance from running water?Be creative, and choose a topic you care about! Your question should be focused; the best projects involve careful and detailed analysis of a simple question, rather than overly ambitious and haphazard treatment of complicated issues.
Are people who smoke more or less likely to buy organic produce than people who don't smoke?
Do student-athletes have better hand-eye coordination that non-athletes?
Can the record of a professional sports team be predicted from its payroll?
2. Submit a more detailed proposal, with a
specific
hypothesis and an
explanation
of how you plan to collect the data (due second Friday).
This
should be
no
more than a page long, but clearly outline the scientific content
and
feasibility
of the project. If you will be conducting a survey, you must
include
the exact wording of the survey questions. You must include enough
details about your data collection methodology that someone could
use them as instructions for
replicating your work.
3. Submit the raw data you collected (due third Friday). Turn this in to me electronically, by emailing a spreadsheet (e.g. from Minitab).
4. Present your findings in class on the
last day of the block, using a PowerPoint (or similar)
slideshow. State your hypothesis, give
descriptive
statistics
and useful charts for your data, describe the statistical
inference
you
employed (i.e. give confidence intervals or P-values), and explain
your
conclusions. Feel free to use your creativity to make your
presentation
eye-catching if you like. However, the most important thing
is that the
viewer should easily be able to determine the questions you asked,
the
methods you used, and the conclusions you drew. These
main
points should jump out at the viewer - use large font!
5. Turn in a paper on the last day of the block,
co-authored by all the members of your group.
It will cover the
same topics as your presentation, but in more detail. The
paper
should be a stand-alone document; do not assume the reader saw
your
presentation. It should include the following sections:
I. Introduction
A brief (<1
page)
description of the general questions you are interested in.
Give
a preview of your most important one or two results ("We will show
that...")
II. Methods
Describe in detail how
your
data were collected. If you conducted a survey, include the
actual survey as an appendix. You must give enough detail
that a
skeptical reader could reproduce your methods exactly.
III. Results
Summarize your data
numerically and graphically. Include enough figures to
illustrate
the main points, but no more than that.
Carry out statistical
inference (hypothesis testing and/or confidence intervals).
This
is the most important part of the paper. Clearly state your
hypotheses, test statistics, p-values, and practical
conclusions.
Be sure to use notation and terminology correctly. Explain
why
the assumptions of your statistical methods are satisfied by your
data.
IV. Discussion
Summarize in
non-technical
terms your major findings.
Discuss the strengths
and
weaknesses of your study. If you failed to find significant
relationships, do you think it is because they don't exist or
because
you lacked statistical power? Were there other flaws in your
methodology?
Overall paper guidelines: Write using the active voice ("We found
that..." instead of "It was found..."). Your writing should
be
succinct and precise, not wordy. Get to the point as
efficiently
as possible. Write in the style of a professional scientific
article - do not discuss your personal feelings about the topic,
describe how you got stuck at a certain stage of the process,
etc.