class: center, middle, inverse, title-slide .title[ # L0: Course Orientation ] .author[ ### Jean Morrison ] .institute[ ### University of Michigan ] .date[ ### Lecture on 2026-01-08 (updated: 2025-12-30) ] --- `\(\newcommand{\ci}{\perp\!\!\!\perp}\)` ## Welcome + Course Basics `\(\newcommand{\ci}{\perp\!\!\!\perp}\)` + We will meet Tuesday and Thursday 5-6:20 pm in SPH2 1123. + Textbook: - ["What If?" By Miguel Hernán and James Robins](https://miguelhernan.org/whatifbook) - Other readings will be provided as pdfs. + Office Hours: - Monday 3-4pm in-person in Jean's office (M4148) or the Conference Room (M4034) - You can request to join by Zoom. I won't run Zoom if there are no requests. - Or by appointment. --- ## Course Structure Tour of [Canvas site](https://umich.instructure.com/courses/809102) --- ## External Course Website - All of the lecture materials and readings are available [on my personal website](https://jean997.github.io/BIOST_881_causal_inference/). - These will remain available after the class indefinitely. --- ## What is Causal Inference? + Causal inference is the art and science of using statistical tools to answer causal questions. -- + Interpreting statistical parameters as causal parameters requires: - A language and philosophy of causality. - A model of the system you are studying. --- ## What We Will Learn in this Class -- + Languages of causality: - Counterfactuals and Potential Outcomes - Graphical models - Structural Equation Models -- + Identifiability conditions: When would a parameter be estimable with infinite data? -- + Estimation strategies: - Weighting, standardization, G-estimation - Propensity scores and matching - Instrumental variable analysis - Methods for time-varying treatment - `\(\dots\)` --- ## About Me - Jean Morrison (call me Jean, Dr. Morrison is also ok) (they/them). -- - Research Interests: + Statistical genetics + Instrumental variable analysis with genetic instruments + High dimensional biological data + Data integration (combining multiple sources of Hi-D data) + Applications of ML tools in causal inference. -- - I like animals and outdoor activities (hiking, biking, kayaking, etc). --- ## About This Class - My goal is to create an active, collaborative, learning environment. - Ground rules: + Be engaged, ask questions, and contribute to discussions. + Complete assignments on schedule. + Be respectful of me and your classmates. - Please contact me if something is not working for you or you are struggling. --- ## Attendance and Recording - There will be no virtual (Zoom) option for attending class remotely - All classes will be recorded and available on the Canvas site. - Recordings will be available only to other students in this course. - In-class activities cannot be made up if you miss class. However, you can miss up to four or 20% of activities (whichever is greater) and still receive full credit. --- ## Due Dates and Late Policy - You are strongly encouraged to meet the deadlines laid out in the course schedule. - However, if something prevents you from meeting these deadlines, you will not be penalized as long as you contact me ahead of time to arrange an alternate deadline. + Please email at least two days in advance (not counting weekends). E.g. if the deadline is on Monday, email me by Thursday. - There are three exceptions to this policy: + The progress report must be handed in on time because it will be reviewed by your peers. Allowing late submissions would not be fair to others who need time to read your work. + Flexibility for the final report deadline is very limited due to end of semester grading constraints. + Final presentations occur on the last one or two days of class. --- ## Intro Survey Please fill out [the introductory survey](https://forms.gle/efUsiyqLV4cR1ZAF9) by Friday 1/9. --- ## Generative AI: Homework - I encourage you to do the homework without using generative AI. + Similarly, I encourage you to do the homework without asking your friends for last year's key. - Trying the homework on your own will help you see what you understand and what you do not understand. + If you are stuck, come to office hours, check the book, and review the lecture notes. - Treat the homework as a learning tool. --- ## Generative AI: Project - Generative AI can be used to help with coding problems in both the project and homework. + I think you will learn best if you try yourself first, find a specific problem, and then ask the AI for help with the specific problem. - It may be useful to talk to a generative AI tool about your project. + Keep in mind that the AI may be wrong and may give you incorrect references. + Often you will get better information over the course of a conversation than in a single prompt. - It may be useful to use generative AI to refine or proof read your writing. + I do not recommend having generative AI write large portions of your report. --- ## Generative AI Statement - Please include a generative AI statement at the end of your project report. Examples below borrowed from Irina's 699 slides: + "No generative AI tools were used for this report” + “I used ChatGPT version XXX in this report. Specifically, I provided ChatGPT with initial drafts of my Abstract, Introduction, and Conclusion and asked it to improve it with respect to flow, grammar, or clarity. I did not use ChatGPT for any assistance in conducting the statistical analysis.” + “I used UM GPT to develop code for the statistical analysis. Starting with the following prompt XXX, I modified the prompt based upon its initial responses and then copied and further modified the code myself. I did not use UM GPT or any other generative AI for writing the report itself.”