The Case for Data Literacy

The worth of data literacy is like the value of a general contractor in a home renovation.

Maybe you’ll need a plumber to redo the kitchen or a structural engineer to add a bedroom – but to successfully complete the remodeling, you need a general contractor to coordinate the right resources and understand the overall state of the project.

At the Government Accountability Office (GAO), general contractor proficiency is the goal for data literacy. Much like when remodeling a home, not everyone needs to be a data specialist to make essential professional judgments – but everyone must be data literate to evaluate the data that impacts their work.

“We’re not trying to make them into statisticians or data scientists,” said Taka Ariga, GAO’s Chief Data Scientist. “We just want to make sure they understand the core concepts, can supervise them accordingly and know what right and wrong looks like when they scope these types of capabilities in their work.”

Without data literacy, employees by definition defer judgment to those with statistical, mathematical or data science specialties – even when they already have domain expertise and experience. This isn’t a best practice, not in GAO’s book. With or without data backgrounds, all employees can possess a basis of assessing the appropriateness of data methodologies, outputs and contexts in their line of work. When they do, they create a data-driven organization.

Who Are the Key Players?

Middle management are the key players.

Junior staff who are “digital natives” may innately understand how to use data for their jobs, but they don’t always understand why they should. Senior leaders may understand why data literacy is foundational to evidence-based policymaking, but they don’t always understand how it is done.

“It’s the middle layer that has to understand not only the why, but the how,” Ariga said.

The challenge, frankly, is that data literacy is not part of official performance evaluations. Already juggling a full load, data literacy can seem superfluous and even daunting for middle managers and staff. Training for maybe later, maybe never.

GAO tackles this challenge by articulating the benefits directly to performance criteria.

It ceases to be a waste of time when staff understand that applying data science will help them execute audits more efficiently, save time and make stakeholders happier.

It becomes less daunting when they understand data skills as an additional professional judgment that gets layered on top of their existing skill sets.

And it becomes vital when they realize data literacy keeps them in control of their judgment calls, not passed over to data specialists.

What Are the Best Practices?

To institute a data literacy framework, it must be in alignment with the agency’s mission.

“We don’t want to treat all data literacy as equal,” Ariga said.

Someone who works in national defense requires different data skills from those in environmental or financial management auditing. “We firmly believe it’s not a one-size-fits-all approach,” Ariga said. Training must be catered to tradecraft. It’s the reason GAO is creating its own data literacy curriculum specific to the oversight community, instead of relying on third-party training that focuses on generic, often commercial aims.

Additionally, the best time for people to learn data skills is when they actually need them.

On-demand tools such as microlearning videos and a walk-in Genius Bar ensure staff can access data solutions and build literacy when they need, instead of waiting months to register for a class.

Hands-on learning is also key. You don’t learn how to ride a bike by reading a user manual. “You ride the bike,” Ariga said.

At GAO, a sandbox construct assures freedom to practice new skills without the fear of failure. Employees may not want to conduct a complex regression analysis for the first time on a big project. But they can do it comfortably in a closed safe space.

What Should We Avoid?

Avoid building a data literacy echo chamber. When one person’s perspective drives the entire framework, it can build in implicit and explicit biases.

“Treating data literacy as a team sport is paramount,” Ariga said. And it’s how GAO is approaching it.

“It’s not ‘Take the Chief Data Scientist’s word as golden,’” Ariga said. GAO’s data literacy framework involves learning center staff who understand how people absorb materials, statisticians who understand the technical mechanics of what is being taught, experience designers who understand the impact of the interactions staff undergo to access training and so on.

This ecosystem of curriculum builders is important because there must be an ecosystem of learners to sustain a data-driven organization.

“The goal of data literacy is not that we train Bob, Suzy and Michael, and they go into their little corners of existence and don’t talk to each other,” Ariga said. “Having that ecosystem of a community of practice is an important sustainment tool. Otherwise, we’re just going to be repeating the same data literacy concepts over and over again.”

Photo credit: ThisisEngineering RAEng on Unsplash