STAT 151A Final Group Project
Fall 2021
The goal of the final project is to apply what you’ve learned in this course to conduct a statistical
analysis. It should be an in-depth regression analysis of a question that interests your group. This
question may come from one of your other courses, your research interests, your future career interests,
etc.
You will work in groups of three to complete the project. It consists of three primary delivarabless:
Project proposal (9% of course grade): due Thursday November 4 at 11:59 PM PST.
Final project report (20% of course grade): due Friday, December 17 at 11:59 AM PST.
Self-evaluation (1% of course grade): due Friday, December 17 at 11:59 PM PST.
1 Data
It is best to start with the question of interest and find relevant data second. As you’re looking for data,
keep in mind your regression analysis must be done in R. Once you find a data set, you should make
sure you are able to load it into R, especially if it is in a format we haven’t used in class before. If
you’re having trouble loading your data set into R, ask for help as soon as possible, so you can make any
necessary adjustments before the project proposal is due.
In order for you to have the greatest chance of success with this project it is important that you
choose a manageable dataset. This means that the data should be readily accessible and large enough
that multiple main effects and interactions can be explored for your model. As such, your dataset must
have at least 100 observations and at least 10 variables (exceptions can be made but you must speak with
me first). The data set should include both quantitative and categorical variables.
Please do not reuse datasets used in examples/homework/labs in class.
2 Data resources
Duke University provides a guide to finding datasets for projects like this, with links to several
possible data sources.
Google Dataset Search allows you to explore 25 million datasets from various sources using search
filters.
The fivethirtyeight package in R contains 128 datasets on topics including politics, sports analytics,
and popular culture.
TidyTuesday publishes a new dataset every week, as well as a list of links to data portals from news
organizations.
This document is adapted from materials created originally by Prof. Elizabeth Purdom at UC Berkeley and Prof. Maria
Tackett at Duke University.
1
The UC Irvine Machine Learning Repository provides 557 datasets and a searchable interface.
Kaggle maintains a large collection of datasets published by users.
If you identify other interesting data sources you are strongly encouraged to share them with your
classmates on Piazza.
3 Detailed timeline and deliverables
Tuesday October 12: project groups formed. You may propose your own groups of exactly
three through this date by signing up here. You are encouraged to use Piazza to share your project
ideas and form groups. Forming your own group is optional. If you do not wish to find your own
group or have difficulty forming one, you may request to be placed in a group by course staff. You
can do this by entering your name on the online form here. Additional groups will be formed at
random from those who submit their names using the form.
Tuesday November 9: project proposals due 11:59 PM. The proposal is a draft of the
introduction section of your project as well as a regression analysis plan and evidence that you can
load the needed data into R. The proposal should be no more than 3 pages (excluding appendices).
Please submit only one document per group to Gradescope, tagging all group members.
You will receive written feedback on this proposal from the professor or the GSI. After writing the
initial proposal, if you do not think your group will work together well this is the best time to come
to me because I can reassign people and form new groups. After this point, it will be difficult to
change the project members.
Friday December 17: final project due 11:59 AM. The purpose of the report is for you
to demonstrate your ability to ask meaningful questions and answer them with the results from
regression analysis, and your proficiency in using R and in interpreting and presenting results.
Focus on methods that help you begin to answer your research questions. You do not have to
apply every statistical procedure we learned. Also pay attention to your presentation. Neatness,
coherency, and clarity will count.
At a minimum, your report should have the following sections: an introduction, a section describing
your final regression model, a discussion section outlining conclusions drawn from the model and
limitations of the analysis, a conclusion summarizing the main findings in answer to the research
question, an additional work section describing other models and diagnostics you tried, and a
references section listing the sources you used. The first four sections should be no more than 3
pages in total, and the “additional work” section should be no more than 5.
At the end of your report, you should include a code appendix containing all the R code used to
conduct your analysis. You may either append the entire .Rmd source, or extract the code chunks
and present them separately. The code appendix will not count toward the page limit. Please
submit only one document per group to Gradescope, tagging all group members.
Friday December 17: self-evaluation due 11:59 PM. Each member of the group must indi-
vidually submit a description of how the work was distributed throughout the project. You should
be specific about how you decided to organize the work and what each group member contributed.
These reports do not need to be formal, but should be about 2 paragraphs.
4 Assessment
The proposal and final report will be graded holistically. Below we provide the rubrics used to do this
grading; for each assignment, each part of the rubric receives an equal weight in the grade. In addition,
2
if the end-of-project self-evaluations show highly disproportionate distribution of effort among members
within a group, grades may be adjusted to reflect this.
Proposal rubric:
Research question and data: Is a research question chosen and clearly defined Is the question
challenging, interesting, and specific Is a specific dataset identified that is relevant for this question,
and can the group load it successfully in R in a usable form
Choice of analysis: Is the model and analysis chosen relevant to the research question Does
it make effective use of the data in hand Is a clear plan given for making modeling choices and
assessing assumptions, with attention to the broader scientific context
Clarity and concision: Is the report organized logically with a clear structure Are individual
steps in the analysis explained clearly Is writing concise and grammatically correct Are length
and formatting requirements observed
Final report rubric:
Choice of analysis: Is an appropriate research question chosen and clearly defined Is the model
and analysis chosen relevant to the research question Does it make effective use of the data in
hand Are modeling choices and assumptions properly justified with reference either to the data
itself (e.g. through diagnostics or exploratory data analysis) or to the broader scientific context
Interpretation and evaluation: Is model output interpreted correctly Are conclusions justified
by specific results from the analysis Are weaknesses of the approach and alternate explanations
or methodologies considered and discussed
Computations: Are computations and calculations performed correctly Are they documented
clearly but concisely Is code (in code appendix) complete and well-structured, using appropriate
naming, formatting, and syntax
Visual presentation: Do plots and figures convey important information that contributes to
the central arguments of the report Are plots and figures labeled appropriately and referred to
effectively in the text
Clarity and concision: Is the report organized logically with a clear structure Are individual
steps in the analysis explained clearly Do conclusions drawn from the data analysis respond
meaningfully to the research question Is writing concise and grammatically correct Are length
and formatting requirements observed
3
