Statistician Katherine Thompson Describes Her Path to the U.S. Census Bureau
February 16, 2016
The American Statistical Association recently spoke with Katherine Thompson, the Methodology Director of Complex Survey Methods and Analysis Group in the Economic Statistical Methods Division of the U.S. Census Bureau about her position as a statistician. Below is an excerpt of the interview originally published in Amstat News.
What or who inspired you to be a statistician?
I am an accidental statistician. When I first entered college as English major, I decided to retake Calculus 1 for the “easy A.” My professor was dynamic and interesting, so I signed up for Calculus 2 with the same professor and took a permanent break from English after declaring math at the end of my sophomore year.
I added operations research, probability and statistics, and data analysis to my schedule my senior year. I liked operations research and probability. I loved statistics and data analysis. More important, I discovered I had some ability. Every data set I studied told a story. Every theorem I learned had an application I could use.
As graduation loomed, my statistics professor talked to me about pursuing graduate studies. The idea of continuing my studies without any real practical experience frightened me. What if I was good at statistics classes, but not at statistics applications? Instead, I decided to seek full-time employment at the U.S. Census Bureau and pursue graduate studies part-time. My undergraduate experience inspired me to become a statistician, but my employment experience convinced me it was the right thing.
What is the most exciting part of your job?
I think the most exciting part of my job these days might be serving as team leader for directorate-wide research projects. Not only do I get to direct relevant research, but also consult with diverse sets of peers of varying experience and expertise. I have learned to use plain language to describe technical concepts to our subject matter experts, who are not statisticians. I have learned to delegate tasks while mentoring teammates. When I was younger, I sought the tangible benefits of my research. Now, I value the intangibles, where my coaching or directions provide leadership opportunities, learning and knowledge-sharing, or innovation from others.
Name a few specific skills you need to do your job.
The most important skills in my job are “reading, writing, and arithmetic.” Reading is absolutely crucial, especially in the area of survey research methods. With an increasingly tight budget climate and decreasing response rates, survey statisticians are responding with adaptive collection designs, alternative imputation methods [process of replacing missing data with substituted values], and increased use of Big Data.
Writing is just as important—not just for publishing, but for communicating with project stakeholders. In planning and execution, I need to get feedback from a statistically literate audience and industry (subject matter) experts who prefer an “equation-free” language.
Finally, applied statistics requires number-crunching. My arithmetic skills are always in play, whether for verifying a colleague’s results, deriving a linearization variance estimator, or validating proofs provided in a manuscript. Programming skills are essential, especially for conducting simulation studies. I am a more-than-adequate SAS programmer these days, but I hope to learn R soon so I can talk the same programming language as the latest generation of statisticians.
What do you do in your job now that didn’t exist 10 years ago?
I’m a little weak on the actual history, but 10 years ago, I didn’t know anyone who owned a smartphone or tablet. Now I don’t know anyone who doesn’t have one! The Census Bureau has kept pace with these emerging technologies by developing mobile apps.
Ten years ago, I never would have imagined you could fit up-to-date information on 20 economic indicators on the screen of a tiny phone, let alone present it in an intelligible way! Now, I can’t imagine not doing it.
Name one or two favorite blogs or books you have read and would recommend to others.
It’s hard to narrow down the list to two books, let alone blogs! I cut my survey sampling teeth on William Cochran’s Sampling Techniques and still consult it frequently. However, I usually recommend Sharon Lohr’s Sampling Design and Analysis, which covers the same content in a more modern way (with real-life examples) and provides both the design-based and model-based perspectives. Once you’ve selected your probability sample and are ready to analyze the data, I highly recommend reading Applied Survey Data Analysis by Stephen Herringa, Brady West, and Patricia Berglund.
What advice would you give to young statisticians just beginning their careers?
Be prepared for some tedium! The interesting projects are often earned by proving you can perform the less glamorous tasks cheerfully and carefully—such as running simulations, examining plots, and preparing summary tables. Look at what others around you are working on and see if you can eventually visualize yourself doing similar work. Ask questions and invite review before you share your work with a larger audience. Be grateful, not embarrassed, when someone catches an error in your work, and pat yourself on the back if the error was caught early. Keep reading. Your education shouldn’t end when you attain your degree. If you can, try to get hands-on experience while you are earning that degree by participating in a consulting program if offered by your school or through an internship. Finally, be open to change. Give everything time, but try to get a variety of experiences.
This is Statistics’ fifth annual Fall Data Challenge, Get Out the Vote, is right around the corner! You and your classmates will have the opportunity to work in teams to apply your statistical skills to real voter-turnout data and provide insights to inspire more people to vote in the upcoming election. The submission window opens on October 19. With contest submissions opening soon, we want to introduce the real-life statisticians, with experience in election…
The 2020 Fall Data Challenge: Get Out the Vote submission window is almost here! In preparation, you can begin reviewing the dataset with your team now. For this year’s challenge, all submissions must utilize the IPUMS-ASA U.S. Voting Behaviors dataset. This rich dataset includes information about voting behaviors in the U.S. over the past 14 years, including 28 variables…