Final Project

There are a number of options available for the final project.

Grading scale:

• Not Assessable
• Needs Revisions
• Meets Expectations
• Excellent

Deliverables

For all projects below, the only deliverable is a “digital artifact” that could be used for the digital capstone showcase. A digital artifact can take any of the following forms:

• A video presentation
• A podcast-style recording
• A set of blog posts

Work with Leslie to determine the most suitable form for your digital artifact, depending on the option you pick.

In order to earn an Excellent, your artifact must be presented in a way that is clear and engaging for a student who has only completed STAT 155. Not every part of the artifact should be easily understood by an introductory student, but they should be able to take away some understanding even from the more advanced material.

Timeline

You/your group will check in with me in-person over Zoom 3 times over the course of the last 2 weeks of class. In order to earn a passing grade, you must meet requirements at all deadlines.

• Deadline 1: Choose your final project option, topic, and group by Friday, October 9 at midnight CST (or the corresponding time on Saturday, October 10 if you are in a later time zone). This will not be an in-person check-in, but please submit your responses on Moodle.

• Deadline 2: Monday, October 12 or Tuesday, October 13. Schedule a discussion time with me on Google Calendar.
  • Data analysis option: Come prepared to talk about your causal graph, your analysis plan, and exploratory visualizations that you have made.
  • Blog post option: Write 500 words for each post and share with me 24 hours before our meeting so that I can give feedback and discuss with you.
  • Advanced topic option: Come prepared to tell me a bit about what you’ve learned on your topic. If you were giving a 15 minute presentation on the topic, this should be about the amount of content you would cover in the first 5 minutes.

• Deadline 3: Thursday, October 15 or Friday, October 16. Schedule a discussion time with me on Google Calendar.
  • Data analysis option: Come prepared to tell me about the models you’ve fit and your interpretations. Share a draft of your artifact so far with me 24 hours in advance.
  • Blog post option: Complete full drafts of both posts and share with me 24 hours before our meeting so that I can give feedback and discuss with you.
  • Advanced topic option: Come prepared to tell me more about new material that is the equivalent of the second 5 minutes of a 15 minute presentation. Share a draft of your artifact so far with me 24 hours in advance.

• Deadline 4: Monday, October 19 or Tuesday, October 20. Schedule a discussion time with me on Google Calendar.
  • Data analysis option: Come prepared to discuss final points of feedback from the previous meeting (deadline 3). Finalize your artifact draft, and share with me 24 hours in advance.
  • Blog post option: Act on feedback from the previous meeting (deadline 3), and share with me 24 hours before our meeting.
  • Advanced topic option: Come prepared to tell me more about new material that is the equivalent of the last 5 minutes of a 15 minute presentation. Finalize your artifact draft, and share with me 24 hours in advance.

• Final deadline: Friday, October 23. By midnight CST, submit the final version of your digital artifact on Moodle.
  • We will not have a showcase/presentation day in class, but instead, I look forward to seeing you showcase your work for the department at either of our two departmental capstone unveiling days! (One will be at the end of Module 2 and the other at the end of Module 4.)

Option 1: Data analysis

Collaboration: Groups of up to 3. Individual work is fine.

Perform a causal analysis on a dataset of your choice. To earn an Excellent, the analysis must contain one major component per team member.

Options for major components:

• Regression analysis for causal effect estimation with a sensitivity analysis for unmeasured confounding
• Inverse probability weighting analysis for causal effect estimation with a sensitivity analysis for unmeasured confounding
• Causal discovery as a sensitivity analysis (Information on the pcalg R package for causal discovery will be added to our website’s appendix.)
• Mediation analysis for causal effect estimation
• Instrumental variables analysis

Resources for finding data:

• Google Dataset Search
• Harvard Dataverse
• Inter-university Consortium for Political and Social Research (ICPSR)
• IPUMS

Option 2: Blog posts

Collaboration: Individual only.

Write two blog posts explaining causal inference ideas to a general audience. The first post should address one of the three topics below.

• Estrogens and uterine cancer example of selection bias: see Graphical Structure of Selection Bias, Exercise 3.
• The smoking-birth weight paradox
• Pick any media item that has interested you. Write a reaction to it / an analysis of it from a causal inference perspective.
  • If you’re looking to explore some media, the Casual Inference podcast is a fun one!

The second post should be a “tour of causal inference” and lead the reader through all topics covered in our course in a cohesive story.

Option 3: Learn an advanced topic

Collaboration: Groups of up to 3. Individual work is fine.

Dig deeper into course topics or learn a new topic. Examples could include:

• Methods for transportability (generalizability) of effects
• Interference
• Details of methods for time-varying treatments
• Specialized considerations for particular study designs

Option 4: Other

If none of these options piques your interest, I’m happy to discuss alternatives with you.