How to help MN counties make data-driven decisions for social programs

Across Minnesota, counties collect data in the normal course of administering human service programs. When someone calls the county in need of food assistance, for example, a  county worker will record information about their family, job, income, and assets. This information is used to determine eligibility for programs like MFIP (cash welfare) and SNAP (food support), monitor program participation, and provide referrals to community organizations.
 
Few MN counties use this data for program improvement.

I'll admit, I was shocked when I learned this. And not just because I love data! Not using existing data is a lost opportunity – even simple analyses can help programs better understand the families that they serve, their needs, and whether these needs are being met.
 
So, why aren’t counties using the data they already collect for program improvement? There are actually a few reasons:

1. Counties in MN lack easy access to aggregate client data for many social programs. Although workers enter data locally, most data actually "live" in state data systems (eg. MAXIS or SSIS). Only a few counties have access to these state systems. When counties without access want to know something about program participants as a whole, they have to make a formal request to state research staff and wait for it to be filled. And because state research staff are busy, there is often a delay.
2. Counties - especially small ones - lack the capacity to use aggregate data for program improvement. Resources are thin and many localities simply do not have the resources/staff to devote to analyzing aggregate participant data. This is less a critique of local governments and more a plea for the state and other funders to help counties build this capacity.
3. Data systems in human service programs are outdated and do not ‘talk' easily to one another. Many systems currently in use were developed decades ago to monitor basic program participation. They weren’t designed to help counties (or the state, for that matter) understand whether or how a program is working. Because the systems developed apart from one another, the data are hard to link. This means that families that interact with multiple programs have to give their personal information to different workers at different times for different systems. This is invasive and inefficient, to say the least. And it makes it hard for those at the local and state level to understand the holistic needs of those they serve.
4. Related to this point, data systems are difficult to learn. True story: A few years ago I tried to learn about a dataset for a major public program, for a project aimed at helping one MN county use that data for program improvement. In my investigation, I learned no codebook (ie. a map to the data) even existed! All information about that dataset literally lived in the heads of various state workers. This makes no sense!

 
Admittedly, some of these problems are easier than others to solve. And to be clear, there are still major limitations of existing datasets - namely, that they are embedded in systems of structural racism and can perpetuate disparate treatment and outcomes by race and ethnicity.
 
But even recognizing these limitations, there is still much that can be accomplished if counties can better understand program participants, activities, and outcomes. Using existing data - with necessary caveats - can enhance efficiencies and minimize the additional burden placed on program participants. 

So what can we do?

• First, the state should make it easier for county human service programs to access aggregate data about their clients.
• Second, funders can invest in technical assistance to build local capacity to use data for program improvement.
• Third, we can strengthen the connection between county governments and researchers/evaluators to complement local capacity to understand and improve human service programs.

Want to know more about using data for program improvement? We'd love to chat! Contact us to learn more.