Teaching Statement

As a researcher, I believe that teaching and learning are two processes complementing each other, because as Heinlein once said, “When one teaches, two learn.” Since statistics is a cross-disciplinary field, communication has become increasingly important, and teaching is one of the best ways to obtain this skill. During my PhD life at Duke University, I have been a teaching assistant for several introductory and graduate level statistics courses. The students learn by applying simple statistical methods to datasets, and I learn how to clearly express my viewpoints.

I also believe that understanding fundamental statistics is as essential as literacy in mathematics since big data has become popular in the information explosion world. I view statistics as a data translation tool -- from numbers to human knowledge, and I sincerely hope this tool can be widely applied to various areas. Therefore, my goal in teaching is to inspire students to learn statistical concepts and applications, allowing them to benefit from massive data.

Statistics consists of theory and applications, so I put equal weight on both parts in my teaching. For the course “Data Analysis and Statistical Inference”, I assisted in the lecture for one semester and conducted a lab section for another semester. During the lectures, I assisted the instructor to conduct clicker activities for students, which involve multiple-choice questions. After students answer each question by using their clickers, a bar plot of how many people selected each option is shown on the projector screen. If the majority of students do not get it correct, the instructor asks them to discuss with each other and re-answer the question. As a teaching assistant, my role was to provide the students a few hints when they get stuck in deciding between two possible answers. In this way, the students receive immediate feedback on their understanding of statistical knowledge.

During the lab sections in the introductory statistics course, the students use the programming software R to work on application exercises, and I encourage the students to apply what they learned in the lecture to the lab problem set. The main statistical techniques used in the lab usually come from the recent lectures, but the students may forget about some definitions of statistical terms, resulting in difficulty of interpreting their own code output. In this case, I write the definitions on the whiteboard to provide a reference to the students, but I also mention that it was covered in a previous lecture and suggest the students review the related class materials. Most importantly, I try to make sure students know exactly what they are doing in their data analysis by randomly asking a few students to verbally interpret their code output.

I teach not only statistical methods but also statistical intuition because having a “sense” of what models can be right saves much time on trial and error. For example, if one gets 0.99 probability of a rare event to happen, obviously something must be wrong. When teaching, I encourage students to see the big picture -- set aside the mathematical calculation for a few seconds and start from the definitions. Errors such as probability greater than 1 occasionally happens, so getting the statistical argument reasonable can save a lot of debugging. An important aspect of learning statistics is being able to identify fundamental problems in statistical design and results, not just accept a statistical argument without question.

In terms of providing feedback, I use the feedback sandwich method to comment on students' work during office hours, and this makes the students feel good in learning. When a student asks me about his/her answer to a problem, I compliment on which part the student worked out correctly, even when the final answer is obviously wrong. Then I point out what should be improved and ask the student to correct it. After the student makes progress, even if it is just a small part, I praise the student for his/her good work again. One student in the “Introduction to Mathematical Statistics” course told me that my warm and encouraging attitude lit up her day, and that the conversation motivated her to understand deeper in statistical theory.

Furthermore, I monitor students' performance and treat each student as a real person. During the lab sections in “Probability and Statistical Inference”, I handed student's assignments back to them in person, so as to learn the students' names. When it comes to grading exams, I was able to identify students who did not do well, and during the next lab, I reminded these people to attend office hours if they have questions. As the semester went on, their grades improved along with their attitude change.

When I teach a statistics course in the future, I plan to set up real-life application problems that demonstrate statistical methods covered in class. The datasets can come from various fields that catches students' interests; this allows them to connect statistics to everyday life. For example, a dataset about baseball team statistics such as at-bats and strikeouts can motivate students to apply linear regression to predict the number of runs for each team. Other than homework problems, I also plan to assign students to work on independent projects, so they can gain hands-on experience from data collection to hypothesis testing. For exam questions, the fact that many students have access to a calculator needs to be accounted for, so I may provide them the code output and ask the students to interpret the model or fill in the blanks. These techniques can enhance students' learning as well as keeping them motivated.

The process of teaching and learning is a lifelong journey. In addition to doing research and taking courses in statistics, I have also participated in the Certificate in College Teaching program at Duke University. The Teaching Triangles activity allowed me to observe how others teach and reflect on my methods. Through the course “Fundamentals of College Teaching”, I learned to manage the classroom and deal with difficult students, which are transferable skills that I can use even in an industrial career. I also took the course “Teaching Statistics” in the Department of Statistical Science and gained insight on statistical education topics such as how to teach introductory statistics students to think with data. This forms a good starting point to better understand what the current challenges of teaching statistics are, and to keep connected with the statistics academia after graduation.