Professor Kelly McConville Reimagines Harvard’s Stat 100 Course

July 20, 2022
Kelly McConville

“The exciting challenge of revamping Stat 100, which is our most general audience intro stats class, was an opportunity that I couldn't say no to,” reflects Professor Kelly McConville, a Senior Lecturer in the Harvard Statistics Department, on her decision to join the Department.  A survey statistician with interests in machine learning and statistics and data science education, Professor McConville arrived in January 2022 from Reed College, where she was an Associate Professor of Statistics.  We reached out to Kelly to learn more about her reimagined course and her new experiences at Harvard to share with the community.  During a lively conversation, Kelly highlighted her passion for course design, teaching, and mentoring undergraduate researchers.

Stats:  What brought you to the Harvard Statistics Department and what are your favorite parts about the Department so far?


Professor McConville: “This is a fun question to think about!  For the past 10 years, I taught in math departments at small liberal arts colleges that would have somewhere between one and four statisticians.  The idea of joining a statistics department, specifically, where I would be surrounded by statisticians and could attend many weekly seminars, was enticing, especially because Harvard has one of the best statistics departments in the country.”


“An aspect that I value – it’s hard to put it into words – is this aura within the Department.  If you want to do something, whether it's teaching a new class, hosting an event, forging a new collaboration, or starting a book club, the Department will do whatever they can to support you and make it happen – I love that!  Different perspectives are valued and encouraged here as part of the Department’s efforts to continue learning and improving; as a new faculty member, I really appreciate that kind of support.”


“Additionally, I value that the Department authentically cares about both research and teaching, especially as someone who is on the teaching end of the spectrum.”


One of the main ways that Professor McConville advanced her pedagogical ideas this spring was by redesigning Stat 100: Introduction to Statistics and Data Science (formerly known as “Introduction to Quantitative Methods for Social Sciences and Humanities”).  In her conversation with us, she highlighted how the course is designed for all students to learn how to effectively analyze data, whether they think they are a “math/stats person” or not.


Stats:  Why is Stat 100 an important course for students to take?


Professor McConville:I bet that all students will end up using data in their jobs or lives to make decisions, but because data is often messy and incomplete, it’s tough to know how to use it to make decisions.  Stat 100 is about helping students to learn how to extract knowledge from messy data; along the way, they acquire good data habits, and they develop their statistical thinking skills.” 

“Because a lot of the students come in with mathematical anxiety, one of my less advertised goals with the course is to dispel the notion that this class isn’t for everyone.  I believe in the growth mindset and want students to finish the class thinking that they're a statistical rockstar (or at least to see how statistics is worthwhile in their daily lives!).”

Stats: What changes did you make to the Stat 100 curriculum?


“The original version of the course had a title that identified the target population of students, but we changed this because I wanted this class to be for any student, regardless of their concentration.  As you would guess from the title, we've also added more data science topics into the course like ‘data wrangling.’”

“Another important change includes taking a more computational route (as opposed to a more mathematical approach) to understand foundational statistical ideas.  A benefit of using a computational approach is that the barrier to entry is a little bit lower; I assume that students have had zero programming experience before the class.  From week one, students use code (I teach Tidyverse Packages in R) to help them understand a new statistical concept.  With these packages, there’s not as steep a learning curve, which helps students to develop their computational skills faster.  We generate, graph, and visualize distributions and look at what happens when we tweak different pieces of the code.  We use computation to simultaneously learn the concepts and how to do statistical analysis.”

Stats:  Was there a specific part of the Stat 100 course experience that stood out?

Professor McConville: “For the first time, I was teaching a class of over 130, so I was worried that I’d struggle to connect with my students and to build a supportive learning community.  However, my Stat 100 TFs/CAs and students did an amazing job of helping me build a fun, engaging environment and made me feel so welcome at Harvard.  My favorite moments were all the little conversations that I had with students about the books they were reading, their career aspirations, the ways they were using stats, or even about a lackluster menu in the dining hall.  At the end of my last lecture, this was also the first time that I had a line of students who wanted to take their picture with me– that was pretty fun!”

Stats: What projects are you working on this summer that you are excited about?


Professor McConville: “To help me with forestry data science projects, I am employing a group of undergraduate researchers.  These projects are directly motivated by researchers at the U.S. Forest Service Forest Inventory and Analysis Program, the U.S.’s survey agency for our nation's trees.  This awesome group of five students are working on different problems in pairs; they are running simulation studies to compare estimators, exploring ways to measure and report uncertainty, and creating interactive dashboards for disseminating the work.  One of our projects centers on producing estimates and uncertainty estimates of above ground carbon, which is important data to know when considering climate change.  Through these projects, I emphasize how to share estimates in a way that is digestible to a non-statistician (to the people who own or manage forests).  We try to think through questions like, ‘how do you give people an intuitive understanding of uncertainty?’  The projects focus on statistical communication for a broad audience.”


Stats: What is something new that you would like to achieve in the next year?


Professor McConville: “I'm excited about designing a new statistical computing course for the spring.  This course [title is still to be determined] will have no coding prerequisites.  Students will progress from learning how to use someone else's code to imagining and creating their own code, all with an eye towards best practices for statistical analysis.”


“The other project that I’m looking forward to is related to undergraduate research.  As you know, I'm a big fan of undergraduate research for many reasons: students learn statistics by doing statistics, they develop communication skills, they strengthen their connection to the field, and they benefit from greater career clarity.  I feel energized by a proposal (from former Chair Neil Shephard) to serve as the Department's undergraduate research catalyst to help students learn about statistics related undergraduate research experiences on and off campus.” 


As we begin the fall semester, we anticipate sharing with the community more about Professor McConville’s new computing course and her initiative to connect more undergraduates with research opportunities.