Last Wednesday, I walked into a classroom at Harvard to teach the very first lecture of my STAT 100 course. My colleague Eric Moskowitz was there to document it for Harvard’s FAS Current.
The first lecture is the opening act of the entire semester. It’s where energy of the course is established, the style of learning is communicated, and students decide whether the course is for them or not. Although I taught introductory statistics and data science courses several times before, making a fresh start at a new place makes one reconsider the whole curriculum.
I spent a lot of time deliberately designing this particular first lecture. Whether you are stepping into the classroom for the very first time as a new professor or TA, or simply looking for fresh ideas to revamp your first-day routine, I wanted to share the outline, activities, and thought process behind how I ran my first class. Needless to say, if it is your first time teaching in a classroom, you should go see the classroom beforehand to see the layout, write with chalk/markers on the blackboard, and use the audio visual setup.
Introductions and Community Building
Learning is a social act. Because STAT 100 relies heavily on neighbor discussions and a team project throughout the semester, social bonds need to start on Day One.
Introducing the Teaching Team: The preceptor and I introduce ourselves and introduce the rest of the teaching team to signal a strong support network. I also share a few lighthearted fun facts on screen to humanize myself and set an approachable tone.
Getting Students Talking: Immediately after, students turn to their neighbors to answer a few quick prompts—like their name, major, and anything else they want to share with their neighbors.
Breaking the ice early ensures that turning to a neighbor during future lectures feels natural from the very start.
Pedagogy as Content: Course Policies Backed by Data
Students often enter intro stats with anxiety about math and programming. I address this immediately: STAT 100 is not a math course, there are no coding prerequisites, and everyone belongs in the room.
Once the anxiety is lowered, I share my pedagogical approach explicitly—using learning sciences publications. I started this trend a few years ago when I realized it is a good idea to teach students how learning works. Also, many statisticians have their one area of application they bring more to class. For instance, some talk more on sports statistics, some more on biostatics. Given my own background in educational research, it made sense for me to bring in more of studies in learning sciences.
Three pedagogical ideas that students meet in the first week is calibrating difficulty, AI policy, and no-tech policy.
Calibrating Difficulty: Students want to know if it is an easy-A class or if half of the class will fail. To explain why the course will be challenging but not overwhelming, I talk about Bjork & Bjork’s work on Desirable Difficulties (Making Things Hard on Yourself, But in a Good Way). The goal is to challenge them enough to maximize retention without making them miserable.
AI Policy with a research study: On Day One, students read the abstract of a study on high school students in China examining the relationship between AI reliance, homework scores, and exam scores. The paper uses core statistical concepts (variables, descriptive statistics, relations between variables, prediction, comparison) while illustrating the real trade-offs of using AI as a cognitive shortcut.
No-Tech Policy via Homework: To support my strict in-class no-laptop and no-cell-phone rule, their first homework assignment includes reading abstract of a study on laptop use and multitasking.
By embedding course policies directly into the curriculum through educational research, students don’t just see arbitrary rules—they engage with statistical reasoning and see the evidence behind how they learn best. Also seeing statistics about learners (which they also are) can be more engaging than seeing statistics about athletes doing a sport that they do not understand.
Technical Content: Reproducibility from Day One
Instructors are often surprised to learn that I introduce R, Quarto, and GitHub in the very first week. Diving straight into technical content sets a clear expectation for the rest of the semester: in my course, data science is done reproducibly, starting on Day One.
Reclaiming Lecture Time: To make room for tech right away, I assign an asynchronous syllabus quiz instead of spending 30 minutes reading policies out loud in class.
Setting High Workflow Standards: I am very opinionated about data science workflows. Introducing tools like Quarto and GitHub immediately signals to students that reproducibility isn’t an advanced add-on—it is the baseline expectation from start to finish.
Scaffolding and Asynchronous Support: High expectations don’t have to mean high anxiety. We take things step-by-step during live instruction, and I record video walkthroughs so students can review the setup process asynchronously at their own pace.
What is to Come: Framing the Discipline and Dry-Running Tools
The first lecture serves as a preview of the entire semester—both conceptually and practically. I use this time to map out what lies ahead while eliminating technical friction before deep coursework begins.
Tracing the Roots of the Field: Before diving into formulas or code, I give a brief tour of how the discipline evolved—from early probability theory in France to classical inferential statistics and today’s modern data science. This historical framing helps students see how the course fits together and what kinds of questions we will tackle as we bridge probability, statistical inference, and data science.
Dry-Running Digital Tools: To prepare students for subsequent lectures, we conduct low-stakes interactive exercises right away. For example, I assign a quick online quiz where students click on a world map to drop a pin where they were the previous week.
Eliminating Future Technical Friction: The real objective isn’t collecting map data; it’s ensuring every student knows how to log in, navigate the interface, and interact with image-based questions. Resolving account setups and UI confusion on Day One prevents technical delays when we analyze complex statistical graphics in future lectures.
Falling in Love with Statistics on Day One
Beyond logistics, software setups, and course policies, the primary goal of the first lecture is inspiration. Many students arrive expecting a dry, mechanical requirement; my job on Day One is to shift that perspective entirely and help them fall in love with the discipline.
What Statisticians Actually Do: I frame statistics not as a set of formulas to memorize, but as a creative lens for asking meaningful, real-world questions about the world around us.
Igniting Curiosity: By showcasing compelling real-world problems that data can help us understand, students see the power and relevance of statistical thinking before we ever dive into code or equations.
If students walk out of the first class seeing statistics as an empowering tool rather than a daunting obligation, half the teaching battle for the rest of the semester is already won.
Note to First-Time Lecturers
If you are stepping into the classroom for the very first time, give yourself grace: teaching is genuinely hard work. Many of the strategies I rely on today—like dry-running polling tools to pre-empt technical hiccups—only exist because I have experienced every possible technical failure over years of practice. Those instincts accumulate over time; you don’t need to have every detail perfected on Day One.
Your first term will feel like a steep learning curve. Teaching can be one of the most fulfilling parts of academic life. Few experiences compare to the energy of being in the room with students, hearing their thoughtful, and watching their curiosity take off. Best of luck with your first lecture—it only gets better from here, and I hope you come to love teaching as much as I do!