[Image courtesy of NBCWashington]
I was traveling last week so I’m late to the party, but I had to share this terrific piece by my friend and fellow election geek Sam Derheimer in the most recent electionlineWeekly. He describes a new effort begun in the District of Columbia to collect data on the election process in order to assist election officials and voters alike. Check it out:
“Research is creating new knowledge.” -Neil Armstrong
During last year’s general election, instead of working the polls at the elementary school across the street, I helped election officials in Washington D.C. better understand Election Day from voters’ perspective.
The D.C. Board of Elections assembled data collection teams from its trusted poll workers to gather information on the voter experience at the polls, specifically to help understand and address delays on Election Day.
The project itself was narrow in scope and methodology — and therefore easily recreated in other jurisdictions — but the potential effect could be transformative.
On Election Day, five teams of three were dispatched to high traffic precincts, where we were stationed for the entire day. We were given instructions, tools, and access to shadow the poll workers, while not interfering with voting.
In hour shifts, we rotated between the check in process, the ballot distribution process, and a flex position whose assignment was to gather data on a more macro level.
Primary Data points
- + Wait time to get to a check in clerk
- + Number of check in clerks working
- + Timing of check in process per individual voter
- + Wait time to get to a paper ballot clerk
- + Timing to distribute paper ballot
- + Number of paper ballot clerks working
- + Wait time to get to a touchscreen machine
- + Timing to set touchscreen machine for voter
- + Number of touch screen machines and number of touch screen clerks working
Total vote time (from entering the check in line to casting a ballot)
- Stop watch
- Printed survey sheets (in case of internet trouble with the tablets)
- Binder with instructions and key information on the project
Prior to Election Day, the data teams received a half-day training that covered both the metrics and the tools. In order for the data to be meaningful, project coordinators needed standards and clear instructions so that the wait times recorded at one precinct were consistently measured with the wait times recorded across town.
During the training session, D.C. officials engaged the team members, most of whom were experienced poll workers and have rich knowledge of the voting process, on how to define the metrics (e.g. when exactly is a voter “in line” for a paper ballot). In that way, D.C. tapped into one of its best resources – its poll workers.
As important as clear data definitions, we were instructed to check in with the precinct captain and poll workers prior to the 7:00 am opening of polls. The precinct captains had been informed of the project, but adding a few roaming data collectors to an unavoidably chaotic day could lead to added stress. Like so many situations, communication can inoculate you from potential problems.
Also key, the recorded timings weren’t linked to the individual poll workers – this wasn’t a measurement of a poll worker’s efficiency, it was a measurement of the precinct’s efficiency. Initially, there was mild curiosity from the poll workers, but that soon evaporated as voters flooded the precinct and they got busy. Then it was time to stay out of the voters’ and poll workers’ way and start timing.
Though, inevitably, in slower times, some poll workers turned around to ask about their own performance. Personally, if they asked, I gave them their last timed transaction, and then tried to move on so as not to influence their behavior. But in the future, it may improve similar projects to establish rules on how to address poll workers who want to know about the data collected.
The data was entered into the tablets and uploaded every hour. Election officials were getting hourly updates on the performance of the select precincts. A broader project could therefore inform resource allocation District wide, as the election progresses through the day.
The data was then compiled into an after election report for the D.C. Board of Elections and will help election officials not only track, but also help alleviate delays and problems on Election Day.
With data on average wait times, by hour, by position at the polls, throughout the entire day, election officials can better allocate resources – both staff and machine. Precinct management can be form fit according to the data to assign poll workers to anticipated choke points in the voting process before they become overtly problematic.
Poll worker duties that are currently performed by two separate positions may be combined where tasks are completed with consistent quickness. Tasks that measured longer may be divided and conquered.
Additionally, the data should inform poll worker trainings for future elections. The better D.C. can educate its poll workers on from where and when delays originate, the better they can apply resources to prevent them from occurring to begin with.
For D.C., the cost for the project was 15 poll workers and staff time to develop the metrics and configure the tablets, which D.C. already owned. But the value of the data is much larger to any jurisdiction seeking to shorten lines and improve the voting experience.
And there was one other set of recommendations that emerged from my data collection team. One of my teammates, a native Russian, believes that American elections would benefit if precincts pumped in music, and even more from a cart offering a selection of teas and cakes for purchase.
So D.C., data is great, but let them eat cake.
It really is gratifying to see more jurisdictions engaging in direct, systematic data collection aimed at studying and improving election administration. Thanks to the District for its interest in such a project – and thanks especially to Sam for taking the time to share it with the rest of us!