My intern experience on the ACLU data science team
When I made the decision to attend graduate school, I did so to pursue my passion for data science for the dismantling of oppressive power structures. The summer after my first year, I searched for internships at organizations that combined these passions with my newly formed skillset, and ultimately landed a role as a data science intern on the ACLU’s newly formed data analytics team. The opportunity was a perfect fit: it allowed me to hone my existing data skills, learn new ones, contribute to a cause I care about, and spend a summer in New York City. What’s not to love?
I was quickly read into two analytics projects. The first was an analysis of the impact of a turnout experiment conducted by the ACLU’s Michigan affiliate in the 2018 midterm election. The affiliate undertook a phone-banking campaign to boost support for Proposal 3, a ballot initiative to amend the state constitution to enshrine straight-ticket voting, automatic voter registration, same-day voter registration, no-excuse absentee voting, and other expansions of voting rights. To measure the impact of the phone calls, the affiliate randomly split its contact list of Michigan voters into treatment precincts, where phone-bank volunteers would call voters in support of the bill, and control precincts, where voters would receive no such phone call from the ACLU. My job was to examine the precinct-level results and determine whether voters who had received the phone calls actually voted in favor of Proposal 3 at higher rates than those who had not.
The task seemed simple enough, but it required weeks of research, phone calls, and trial and error to accurately gather, clean, merge and analyze the data. Ultimately, the analysis concluded that there was not sufficient statistical power to assert the phone calls had a causal effect in persuading voters to support Proposal 3, but fortunately, the proposal passed anyway with 67% of the vote.
My second project was to design a “list puller” that development and political staff could use to create lists of ACLU members and supporters based on specified criteria, and would ultimately lighten the analytics team’s lift when receiving list pull requests. I worked to scope the project by meeting with stakeholders across the organization, designing a new data schema and building out a full-fledged mock-up of the tool using R’s Shiny package (a skill I learned for the project!). This project, while less structured than analyzing the results of the Michigan experiment, allowed me to learn a new skill, work with many different folks across the organization, and ultimately create a prototype that would serve as a foundation for a useful tool that will be fleshed out in the months to come.
My two assignments outlined just how much diversity there is in the types of data tasks the team takes on — the Michigan project was more analytical, with a lot of data munging and statistical thinking required, while the list puller was much more open-ended with opportunities for creative design mixed in with the technical requirements of joining disparate data sources together. I particularly enjoyed getting to work with analysts on different parts of the team, namely “Advolytics” (advocacy analytics) for the Michigan experiment and “Engage-alytics” (fundraising & engagement analytics) for the list puller. My professional experience at the ACLU was wholly positive and it encouraged me to pursue career opportunities in political analytics and advocacy.
I also want to shout out the entire ACLU analytics team for fostering a fun, welcoming and inclusive environment and making me feel like a valued member of the team even though I was only around for ten short weeks. In particular, I learned so much from Jung-Hee and Brooke and I know I’ll carry the lessons I learned from them forward into my career. As I think back on the summer, there was a lot of hard work, but there were also birthday celebrations, coffee chats, team happy hours, and bubble soccer. I had such a blast, I’m so grateful for the experience, and I’m also deeply heartened to know that there’s a team of such smart, thoughtful, and dedicated ACLU data nerds working to make America a better place.
