Project Description - Statistics 407/507 - Fall 2022
Collect a research-type data set and analyze the data using methods that
are covered in our Experimental Design course. You may use any of the
experimental designs we discuss in class except completely randomized
designs from Chapter 3. Your project is due in two parts. The first part
is just some sentences about the data you have in mind so I can help you
set up the analysis. This will be due on October ? via Canvas. If you have
not gotten a data set by then just let me know and I can help you find a
data set and project. The second part is turning in the project during
finals week, and is due (via Canvas) by Thursday, December 15. For Fall
2022 the project will be turned in via Canvas as a Powerpoint file. An
example of a previous student project is posted on the course Canvas site.
Your presentation should address the following issues (as
appropriate):
- The context of the experiment
- General information about the topic.
- The general hypothesis of interest.
- Specify H0 and H1 for the tests of major interest.
- The goals of the experiment, i.e., what it would contribute to
knowledge in the field. Also, relate it to other experiments already
done in the field.
- Define the experimental design.
- State the design used, and describe the experiment in enough
detail to identify the design.
- Defend the choice of the blocking variable(s) and the levels used
in your experiment. If you don't use a blocking variable, defend why
not.
- Define the treatment factor(s), whether they are fixed or random,
and the levels. Defend why they are used in your experiment.
- Describe the dependent variable, and specify any potential
problems one might encounter in measuring this dependent variable.
- When applicable, use power calculations to calculate the sample
size needed to detect differences that you believe are of practical
significance.
- Describe the population from which the subjects will be selected.
Describe the sampling procedure and the random allocation of subjects to
treatments (if you have an observational study then just describe how
data were collected).
- Describe and conduct the data analysis.
- Write out the Linear Model using notation like what is used in the
text (not just computer code).
- Lay out the appropriate skeleton ANOVA Table (the Source and df
columns).
- Discuss any uses of post hoc comparisons that may be appropriate
to your experiment.
- Analyze the data. Conduct exploratory analyses and generate some
graphs or plots to describe the data. Present the main results from
the analyses .
- Discuss the implications of your results for this field, and compare
your results to other published results. (Most of this should follow
from your results and the background information in part 1)
These first part (an idea of the data and questions to be addressed) is
due by October ? on Canvas.
Special thanks to Dr. Bill Mickelson, whose Stat 401 Project Description
I have borrowed and adapted to make this Project Description.
Note: the rubric I use to grade the presentations uses the following
categories: Was the motivation for the research clear? Was the process of
data collection clear? Was the choice of experimental design clear? Was
the choice of experimental design appropriate? Was the data analysis
explained clearly? Was the data analysis appropriate (Including graphs)?
Were assumptions checked? Quality of overall presentation and
effectiveness of communication?