Most families who need support are not missing out because the help isn't there. They are missing out because finding it, qualifying for it, and asking for it takes more time and energy than they have. We are trying to measure exactly how much.
Navigation burden is the work a family has to do to turn help that exists on paper into help that actually arrives. Every phone call. Every form. Every office that says you need to talk to a different office. Every document you have to find, copy, and mail.
It is easy to count how many programs a county has. It is much harder to count what it costs a person to use one. That cost has been studied carefully, but almost always one program at a time. A researcher takes SNAP, or Medicaid, or a disability claim, and measures what that one application requires of the people who apply to it (Homonoff and Somerville, 2021; Herd, DeLeire, Harvey and Moynihan, 2013; Deshpande and Li, 2019). Those findings are reliable. They describe only that one program.
A real search involves more than one program. Looking for help might start with a phone number nobody answers, then a county office that gives out a second number, then a form that requires a birth certificate you have to pay to replace, then a food pantry open only during the hours you work, then a nonprofit that serves the next county over. At some point in that sequence a person decides whether to keep going. That decision is made against the total, not against any single program's share of it. The existing evidence already points here: reviewing the take-up research, Currie (2006) found that removing individual barriers does not necessarily have much effect, and concluded that one must address the whole bundle.
No existing records add that total up. Each program keeps records on its own applicants and nobody else. The calls that produced nothing were never anyone's case file. A person who hung up and did not call back does not appear in any denial rate, because there was never an application to deny. Studies that do reach eligible people who never applied have to be built on records complete enough to find them. Bhargava and Manoli (2015) could study non-claimants of the earned income tax credit only because the IRS could identify them from tax filings, and even then the best explanations were confusion and informational complexity rather than stigma. Where no such record exists, those people are simply missing from the count (Ko and Moffitt, 2022).
The result is that the number deciding whether a family gets diapers this month is the one number nobody has collected. The need itself is documented: in a sample of 877 low-income urban mothers, almost 30 percent reported not having enough diapers, and reporting diaper need was associated with reporting mental health need (Smith, Kruse, Weir and Goldblum, 2013). What is missing is the cost of trying to fix it.
Measuring a full search means asking while it is still going on, and that is the difficult part. We run a resource directory and a mutual aid board, so families reach us during a search rather than months after one ended. That is when the questions are worth asking. A researcher usually arrives months later.
We build on administrative burden, a framework from public administration research (Moynihan, Herd and Harvey, 2015; Herd and Moynihan, 2018) that breaks the cost of using a public program into three parts. We are extending it from government programs into the wider world of family resource-seeking, where a lot of the help comes from nonprofits, churches, and neighbors instead of agencies, which is now where much of the assistance actually sits (Allard, 2009).
Finding out that help exists, and whether it applies to you
In the administrative burden literature these are information acquisition costs: learning a program exists, determining eligibility, and identifying where the process begins. In plainer terms, nobody hands you a list.
Why we start here: this is the cost our own directory is built to reduce, which makes it the one category we can test ourselves against. If putting the right county page in front of someone does not change what they do next, we should be the first to know.
The work of proving you qualify and finishing the process
The costs of complying with a process once you have found it: documentation, verification, time, and travel. This is the category most people picture when they think of red tape, and it is the easiest of the three to measure, because most of it leaves a paper trail.
Why it matters: the same benefit costs different families very different amounts. Being eligible for something is not the same as being able to get it, and most of the distance between those two things is compliance cost.
What it takes out of a person to ask
The costs to a person's sense of self: stigma, stress, and loss of autonomy. Herd and Moynihan treat this as its own category rather than a side effect of the other two, and we think that is right.
Why it is the hardest to measure: it leaves almost no trace in administrative data. Agencies record applications, denials, and approvals. Nothing records that someone hung up because they could not face explaining it again, or never called in the first place. This is the category most likely to be underestimated by the very systems trying to reduce it.
A note on separating them. In real life these run together. Someone who has been through a humiliating intake once carries a higher psychological cost into the next attempt, and may respond by never learning about the next program at all. The current questions ask only which part was hardest, which tells us how people weigh the three against each other rather than how much each independently contributes. Separating their independent contributions is part of what the full instrument is being built to do. Disentangling that properly needs a design we have not built yet, and it is one of the things we would want a collaborator for.
These six questions are a pilot, which is how a good instrument gets built. They sit at the end of our donation form, and their job is to sharpen what the real instrument should ask. They are not a sample, so nobody should draw conclusions from them, us included.
What they do establish is the design principle we intend to carry forward. Every question points at one specific episode: think of a particular time you looked for family resources, then answer everything about that same time. If answers drift across different occasions you cannot connect what was hardest to how many places someone had to call, and the whole thing collapses into impressions.
We ask what people did, not what they needed. "Have you ever looked for family resources" gets an honest answer. "Have you ever needed financial help" makes people flinch, and the families most worth hearing from are the ones who go quiet first.
The real instrument is in development. It takes the questions above as a starting point and then does the things a donation form cannot:
When it opens, it will be linked from this page and anyone will be able to take it. Until then, treat everything here as a description of a design rather than a report of results.
This is one specific comparison the study is being designed to support, not the whole of what we are after. We mention it because it is unusually tractable: the groups are already distinguishable in our data, and the mechanism is easy to state.
When a grandparent, aunt, or family friend takes over raising a child, the child's needs do not change. The paperwork does. Eligibility rules, custody requirements, and proof-of-relationship rules are generally written for legal parents, and a caregiver who is not one can spend months on questions a parent never has to answer.
If navigation burden does turn out to be measurably higher for kinship caregivers, it is a specific and fixable problem, and it points at rules that could be rewritten rather than at families who ought to try harder. That is the kind of finding that changes something, which is why the caregiver question was built to allow the comparison.
These are intentions, not work in progress. We are listing them because the right collaborator usually turns up around a question, not around an organization, and because we would rather be told early that one of these is a bad idea.
Everything we currently measure comes from people who searched. The families who needed help and never started are invisible to us, and they may be the ones the system fails hardest. Reaching them means asking somewhere other than a resource website, which is a study we cannot run alone.
The same request should not take more work in a county of nine thousand than in a city. We suspect it does, through travel, thinner staffing, and fewer offices. Our directory covers both, so the comparison is available to make.
Our directory is a county-by-county picture of what exists. Turned around, it is a map of what does not: which categories thin out where, and whether the gaps line up with the places families report the longest searches.
If learning costs are as large as the framework suggests, then putting the right county page in front of someone should change what they do next. That is testable, and it would tell us whether tools like ours actually reduce burden or just move it.
Most research pages go up once the work is finished. Ours is going up while the work is being built, which means the design is open to scrutiny from the first question rather than after the fact. Today that means a framework we have committed to, a set of preliminary questions live on our donation form, and an instrument in development. Every result we publish here will carry a date and the number of people it came from.
The same openness applies to the weak spots. The questions running today are preliminary by design, and their job is to sharpen what the real instrument should ask rather than to produce findings. When the full survey opens, whatever its limitations turn out to be will be listed here next to its results, with the same prominence.
We are a small rural foundation, and we bring things a university lab usually cannot assemble on its own. Here is what is on the table.
Our published resource directory covers 3,163 counties across all 50 states and the District of Columbia, spanning utilities and financial help, childcare and early education, food assistance, diapers and baby supplies, maternal and infant health, healthcare, housing, and mental health. That is a usable picture of what exists where.
The instrument is in development rather than finished, which is the useful moment to join. The design principle is set, the caregiver comparison is built in, and there is still room to add what matters in your county before anything is fielded.
We are based in Caldwell County, Missouri, and rural families are chronically underrepresented in this literature. We can reach people who are hard to reach from a campus.
If you run a health department, a community action agency, a hospital, an extension office, or a university lab, there are a few obvious ways to work together. Field the study with the families you already serve once it opens. Shape a question with us that matters to your county. Take the aggregate results back for your own reporting and grant applications. Or write it up together.
We are also happy to be the smaller partner. If the work gets done and families are better off, we do not need our name first.
And you do not need a project to talk to us. If you want to sit down for half an hour and compare notes on what you are seeing in your county, we will take that meeting gladly, in person, by phone, or on video. Some of the best questions on this page came out of conversations that started with no agenda at all.
If what you need is a specific piece of work done on a deadline, whether that is a survey designed, a needs assessment run, or an external evaluator named in a grant budget, that is a different conversation and we have laid it out separately on work with us.
The framework we use and the studies we describe are other people's work. They are listed here so anyone can check what we are claiming against the source.
Citing this work does not mean these authors are involved with us or endorse anything on this page.
Tell us what you are seeing in your community and what would actually be useful to measure. We will tell you honestly whether we can help, and we are glad to meet just to talk.
