A Robust Vision for Using Data in Government
State and local decision-makers are learning to harness it as a strategic asset
Data is a strategic asset that can help policy-makers and officials manage programs more effectively and inexpensively.
State and local governments collect data in amounts that almost defy description -- everything from health outcomes and marriage records to crime statistics and Head Start enrollments. Much of this data collection is required by federal or state law, and some is compiled at the initiative of individual agencies.
But collecting data is not the same as harnessing data, a fact that is gaining acceptance among state and local decision-makers as they move toward a more robust vision for what data can accomplish -- specifically, toward the idea that data is a strategic asset that can help policy-makers and officials manage programs more effectively and inexpensively.
Many states and localities are finding innovative approaches to sharing, matching and using data to drive policy decisions, budgeting and operational decision-making. To fully understand the extent to which this is happening, The Pew Charitable Trusts has initiated a rigorous project that will help identify success stories that other governments can study, duplicate and even improve upon.
One way that states can harness information for a larger strategic purpose is to take a deep dive into data that have been collected for years in order to uncover systemic failures and craft new solutions, which is what Delaware's Department of Education did in 2012.
After an extensive analysis of data measuring the performance of Delaware high school students, conducted with the assistance of Harvard University's Strategic Data Project, the state determined that a large number of students whose SAT scores indicated they were capable of obtaining a college education nonetheless were not enrolled in college. From 2008 to 2011, for example, 18 percent of Delaware students who scored at least 1550 out of 2400 on the SAT did not enroll in college.
With these data in hand, the state began what it called its "Getting to Zero" campaign, designed to take that 18 percent down to zero by having every college-ready student in Delaware apply for and enroll in post-secondary education. The campaign includes better training for school counselors on how to complete the Free Application for Federal Student Aid; the designation of October and November as college application months, during which students receive help with college applications; and a texting system that families can use to access real-time information on financial aid and other concerns.
The campaign is proving to be a success. Ninety-eight percent of the state's college-ready applicants from the high school classes of 2014 and 2015, the first two years of Getting to Zero, enrolled in an institution of higher learning.
New Mexico had a different problem to solve: the improper payment of unemployment-insurance claims. With support from the federal government, the New Mexico Department of Workforce Solutions, along with several other state workforce agencies, modernized and integrated its unemployment insurance tax and claims system, significantly decreasing fraudulent payments.
But the state wanted to do more to predict and prevent overpayments. Working with a private-sector partner, state officials mined and analyzed years of data on employment by industry, prior claims, work history and other variables to uncover trends and patterns in the behavior of unemployment-insurance applicants. The data helped state officials predict where in the application process claimants were likely to provide inaccurate responses that could lead to improper payments. Armed with this information, state officials created and tested personalized messages designed to prompt claimants to provide accurate information. They found, for example, that claimants did not change their behavior when the message described the relevant law and penalties for breaking it. However, when told that nine out of 10 people in their home counties reported their earnings accurately, a quarter of claimants were more likely to report their incomes correctly.
New Mexico continues to do randomized control trials to test the effectiveness of the predictions -- updating and recalibrating as needed. The department expects to see a 35 percent reduction in overpayments for a savings of $1.9 million annually.
As states and localities face common challenges, the strategic use of data -- which too often sit underutilized on computer servers and in paper files -- provides an opportunity to not only improve program management but also to share innovations that other governments can use to better serve their constituents.
This piece was previously published in Governing.