No national count of untested sexual assault kits exists — no federal law requires anyone to produce one

What is happening?

The United States has no national count of how many sexual assault kits sit untested. A systematic attempt to build one, through public records requests to all 50 states, documented at least 25,000 untested kits for 2022 and obtained usable data from only 30 states and the District of Columbia. The national figures that circulate instead are extrapolations — a 2018 study put 200,000 untested kits in the custody of police departments, and a 2021 study built from 911 counties in 15 states estimated 300,000 to 400,000 unsubmitted kits accumulated between 2014 and 2018.

The gap is not a research gap. There is no federal law requiring the tracking or testing of rape kits, and there is no national system for collecting and tracking kits held by law enforcement or crime labs. Reform has spread across the states — more than 45 have enacted rape kit legislation since 2002 and Maine became the last state to do so in 2026 — but the counts that result are published in different years under different definitions and are never summed into a national figure.

Whose problem is this?

The body that could require a national number is federal, and the bodies that physically hold the kits are municipal. The party that could order the count never touches a kit, and the parties holding the kits are under no obligation to report upward.

Where does this problem end?

This is not a dispute between rival numbers. Nobody has produced a number for others to disagree with, because no institution has been assigned the job of producing one.

What is the state now, and what should it be?

Now

Needs a new measurementThe target state: no source read here names an official national target for producing a count. There is no federal reporting deadline, no designated compiling agency, and no stated figure that a national inventory is supposed to reach. The advocacy demand is that every state hold a comprehensive record of the number and location of all untested kits, which is a state-level condition rather than a national one.

How big is it?

The affected population is people whose sexual assault kit has gone untested. The published range runs from 25,000 to 400,000, and the width of that range is the finding rather than a margin of error.

The low end is what has actually been documented. A records-request project covering all 50 states found at least 25,000 untested kits held by law enforcement agencies and crime labs in 2022, and 20 states supplied no usable data at all. The high end is an extrapolation from a sample. A 2021 study used data from 911 counties in 15 states to estimate 300,000 to 400,000 unsubmitted kits accumulated between 2014 and 2018.

One kit is treated here as one person. That assumption is not verified in any source read for this round, and it fails in both directions. A survivor can have more than one kit over a lifetime, and some kits belong to cases where no survivor was ever identified.

Under what conditions does it arise?

Four conditions hold this state in place, and none of them is a shortage of money or of technical capacity.

1. No federal mandate exists. There is no federal law requiring the tracking or testing of rape kits, so no agency can be held to a reporting deadline and no agency is in default when the number is missing. 2. The kits are local and the count would be national. Kits sit in thousands of municipal police property rooms and in state and local crime labs, and no channel obliges those holders to report a figure upward to anyone. 3. The definitions differ. A kit can be untested because it was never submitted to a lab, or because it was submitted and then waited in a queue. Jurisdictions that publish counts do not all draw that line in the same place, so the published figures are not additive even where they exist. 4. Money moves without a count attached to it. Federal spending on the backlog has exceeded 1.3 billion dollars since 2011 and the federal SAKI program has inventoried more than 205,000 kits, yet neither the appropriation nor the program yields a published national total.

What has been tried?

Every attempt above counts kits inside one jurisdiction or one program. None of them was designed to produce a number for the country, and adding them together would double-count any kit that appears in more than one of them.

What was found?

Why is it still unsolved?

Measurement absent — no institution is required to produce the number, so the number does not exist.

The usual shape of an unsolved problem is that the size is known and the remedy is contested. This one is inverted. The remedy is barely contested at all. Inventory the kits, put them into a tracking system, publish the result. More than 45 states have now legislated some version of that. What is missing is the layer above the states. No federal law obliges anyone to report, and no federal body compiles what the states already publish.

So the national figures in circulation are of two kinds, and neither of them is a count. The first kind is an extrapolation from a sample, like the 2021 estimate of 300,000 to 400,000 built from 911 counties in 15 states. The second kind is a partial tally that names its own gaps, like the 25,000 kits documented for 2022 across 30 states and the District of Columbia. Both are honest about their limits. Neither can be converted into the other, and the distance between them is roughly sixteenfold.

A count that nobody owes is a count that nobody produces. Maine makes the mechanism visible. Reporting from Maine in May 2026 put it plainly — ask anyone in the state how many sexual assault kits are sitting in law enforcement lockers, and no one can tell you. Maine was the last state to enact rape kit reform, and its tracking system is scheduled to launch fully by 2028. The other states reached that point earlier, and the country still has no total, because the state counts were never built to add up.

What observation would mean it is solved?

Candidates — (a) a federal or federally funded body publishes an annual national count with all 50 states reporting (b) every state operates a tracking system that publishes a current public count on a shared definition (c) the spread between the highest and lowest national figures narrows toward a single number.

(a) alone can be satisfied by a number that is assembled but not trusted. A national report that accepts whatever each state chooses to submit would produce a total on schedule while the definitions underneath it stay incompatible.

(b) alone measures systems rather than kits. More than 30 states already have a tracking system implemented or in progress and the country still has no total, so counting the systems is not the same as counting the kits.

(c) alone can be satisfied by the wrong mechanism. If the extrapolated estimates simply stop being published, the spread collapses toward nothing without a single additional kit having been counted.

None of the three is sufficient by itself, and the first is the only one that would produce a number the other two could be checked against.

What is it connected to?

Fills with researchplausibly connected to crime lab capacity and DNA testing turnaround, to statute of limitations reform, to the federal grant structure that funds forensic backlogs, and to the wider question of what the United States does and does not count about sexual violence. Relation type and evidence grade were not researched in this round.

What these sources do not say

• Which 20 states failed to supply usable data to the one systematic national tally, and what stopped each of them from doing so.
• How many of the more than 30 state tracking systems publish a current public count, as opposed to giving an individual survivor a status lookup for one kit. The two are different products and the sources read here do not separate them.
• Whether the more than 205,000 kits inventoried under the federal SAKI program overlap with the kits counted in state inventories. That overlap is exactly what would have to be known before any of these figures could be added together.
• What a kit means in each state count. No source read here states a shared definition that the state figures obey, and the two standard definitions of untested differ by whether the kit ever reached a lab.
• Why the same organization now publishes a smaller number over wider coverage. The Congressional Research Service records nearly 90,000 kits documented in 37 states and DC as of January 2022; the current page of that same project states nearly 40,000 across 41 jurisdictions and carries no date. The two coverage figures are not published in the same unit, so the difference between them cannot be taken exactly, but the count fell by more than half, and nothing read here says whether that reflects kits since tested, a redefinition of scope, or a page left to go stale.
• What has happened since 2022. Every national figure that this round both opened and could date rests on data ending in 2022 or earlier. Two figures that may be more recent carry no date at all, the nearly 40,000 on the current Accountability Project page and the more than 205,000 inventoried under SAKI, so neither can be placed on a timeline. Press reporting from mid-2025 carrying a further nationwide tally was located in this round but could not be opened, and it is listed in the evidence table without a confirmed URL. No source read here explains why no dated national tally has been published for any year after 2022, nor whether any federal bill to require a national count has been introduced in the meantime.

See the evidence

No original crime lab, police property room or state audit record was read directly. The Congressional Research Service report was read through the EveryCRSReport mirror because both official hosts returned HTTP 403. Two independent sources state the same core fact and were confirmed against each other: the Congressional Research Service in 2022 and the End the Backlog project of the Joyful Heart Foundation as retrieved in 2026 both say that the national total of untested kits cannot be established, and both attribute that to the absence of a counting system rather than to any dispute over the data. The national figures do not agree with one another, because they measure different things. 25,000 is a partial tally of a stock in 2022, 200,000 and 300,000 to 400,000 are extrapolations to the whole country from samples, and more than 205,000 is a cumulative program count of kits located since 2015. The SAKI metrics page carries no as-of date, so its figures are reported here as retrieved on 2026-08-07 rather than as of any stated period. This is a Path A output, so observation_refs is empty and provenance_mode is press-derived.

People affected

Estimated range 25,000–400,000 As of 2014~2022

Sensitivity This interval is not a confidence interval. Its endpoints come from two incompatible measurements and the sixteenfold width between them is the substance of this problem rather than statistical uncertainty around a known quantity. The low end is a partial tally of a stock at one date across 31 reporting jurisdictions; the high end is a sample-based extrapolation of an accumulation over five years across the whole country. Neither endpoint can be converted into the other, and no source read in this round reconciles them. Both endpoints also predate 2025, because no national tally carrying a date later than 2022 was opened in this round. Undated cumulative figures exist, and press reporting from mid-2025 carries a further nationwide tally that could not be opened, so neither could serve as an endpoint here. One kit is treated as one person, which is unverified and fails in both directions: a survivor may have more than one kit over a lifetime, and some kits belong to cases where no survivor was ever identified. A separate cumulative figure, the more than 205,000 kits inventoried by the federal SAKI program since 2015, sits inside this interval but measures a third thing again, namely kits located by one grant program rather than kits currently untested nationwide.

Regional breakdown State-level untested kit counts are published on the End the Backlog state pages, but they carry different years and different definitions from one another — the Colorado value is dated 2026 while the Maine value is dated 2018. The one systematic national collection got usable data from only 30 states and the District of Columbia, so 20 states have no comparable value at all. Summing the published state figures would therefore not reconstitute either endpoint of this interval, and apportioning a national estimate by state population is prohibited proportional allocation with no basis here, since kit accumulation follows reporting and submission practice rather than population.