What happens to a complaint against a New Jersey police officer

When somebody files a complaint against a police officer in New Jersey, the department investigates itself and records what it found. Those findings are published in aggregate, one year at a time, and they are the closest thing the public has to an answer to a simple question: what actually happens to complaints?

Across the 2021 to 2025 reporting years, New Jersey agencies recorded 82,507 dispositions.^[1][2] Here is how they break down.

The largest category is an absence

Disposition	Count	Share
Not reported	22,181	26.9%
Sustained	18,589	22.5%
Exonerated	13,871	16.8%
Not sustained	11,314	13.7%
Unfounded	8,503	10.3%
Administratively closed	8,049	9.8%

Those counts merge the capitalisation variants agencies filed under. The same outcome arrives as both Sustained and sustained, and as Not sustained, not sustained and not Sustained, depending on who typed it. Left unmerged the categories fragment; merged, they are the figures above.

More complaints arrive with no disposition recorded than with any single actual finding. Not sustained, unfounded, exonerated, administratively closed, sustained: each of those is a real answer an investigator reached. The largest bar on the chart is none of them.

That is worth sitting with. Roughly one in four times the state's complaint data is asked what came of an investigation, the record does not say.

Some of that gap is explainable. Agencies changed what they were required to report partway through this period, and a complaint filed late in a year may still be open when the annual figures are compiled. But the gap does not shrink to nothing on either explanation, and it is not evenly spread: it is a reporting practice that varies by agency, not a fixed property of the data.

What the words mean

The four non-sustained outcomes are not interchangeable, and the differences matter.

Unfounded means the investigation concluded the incident did not happen as described. Exonerated means it did happen, and the officer's conduct was found to be within policy. Those are different answers to different questions, and lumping them together as complaints that went nowhere loses the distinction: an exonerated complaint is a finding that the conduct was permitted.

Not sustained means the evidence was insufficient either way. It is not a finding of innocence and not a finding of fault. Administratively closed typically means the process ended without reaching the merits, for instance because the complainant withdrew or the officer left the agency.

Sustained means the agency found the allegation supported.

Only the last of those produces discipline, and only a small fraction of sustained findings produce discipline serious enough to be published with an officer's name attached.

Two sustained counts that do not agree

The Attorney General's releases carry the sustained figure twice, and the two do not match.

One is the disposition breakdown above, which records 18,589 sustained outcomes. The other is a separate per-agency field counting sustained investigations, which totals 14,345 across the same years and the same agencies. The gap is 4,244, or about a quarter of the smaller figure.

A plausible reading is that the two count different things: one investigation can carry several allegations, so an investigation-level count and an allegation-level count would legitimately differ. That reading is consistent with the direction of the gap. But nothing published with the data says which field is which, so it remains a reasonable inference rather than a documented fact.

This site uses the investigation-level field wherever it computes a sustained rate, and says so. The important point for a reader is narrower: a sustained rate quoted from New Jersey data can be either 17.4% or 22.5% depending on which published column it came from, and neither is wrong.

The funnel

Put the numbers in sequence and the shape of the system becomes visible.

Agencies opened 82,507 investigations across these five years. Between 17.4% and 22.5% ended in a sustained finding, depending on the field. Of the discipline that followed, only the most serious becomes public with a name attached: a termination, a demotion, or a suspension of more than five days, and only after every appeal is finished. That set contains 3,006 records across six reporting years.

So the published, named record is a narrow slice of a much larger process, and the narrowing happens at every stage. This is not a criticism of the data. It is the design: New Jersey publishes final, adjudicated outcomes precisely because unadjudicated ones would name people whose cases had not concluded. But it does mean that reading the named records as a full account of police misconduct in New Jersey would be a mistake in both directions. It excludes complaints that were sustained but drew lighter discipline, and it excludes everything that was never sustained at all, including complaints that were true but unprovable.

What this does not tell you

A high sustained rate at one agency and a low one at another does not rank them. Sustaining a complaint is an act of classification as much as an act of investigation, and agencies classify differently. Some record a single incident as several allegations; some record it as one. Some close matters administratively that others carry through to a finding.

Nor does the disposition mix say anything about any individual complaint. An unfounded finding in this data is one agency's conclusion about its own officer, reached without an outside adjudicator in the overwhelming majority of cases.

The honest reading is narrower than it looks and more useful than it sounds. This data describes how New Jersey police agencies process complaints against themselves, at what volume, and with what recorded outcomes. It is a picture of a process. Treating it as a measure of how much misconduct occurs would require assuming the process catches everything, and nothing in the record supports that assumption.

Where to look next

Every agency's own disposition mix is on its department page, alongside the complaint categories those investigations fell into. The allegation categories are compiled statewide, so the same question can be asked of a specific kind of complaint rather than of the whole pile.

The uncategorised share is worth watching in future releases. If the not-reported column shrinks as agencies settle into the current requirements, the data gets more useful every year. If it does not, that is itself the finding.