Most police investigations are not thorough by any rigorous evidentiary standard. Clearance rates, the primary metric departments use, count arrests, not convictions or factual resolution. The research record is consistent: most cases are solvable only when a witness identifies the suspect at the scene. Detective follow-up, in the absence of that initial information, produces arrests in a small fraction of cases. The structural gap between how police investigate and how the public understands investigation is large, documented, and rarely discussed in official crime statistics.
The question of whether police investigations are thorough is not an abstract one. It determines whether the right person is arrested, whether the evidence holds at trial, whether victims receive any resolution, and whether the institution of criminal investigation functions as designed or as theater. The research literature answers the question directly, and the answer is uncomfortable for official statistics.
The Clearance Rate Problem
The FBI’s Uniform Crime Reporting program, which aggregates crime statistics from participating law enforcement agencies across the country, measures investigative success through clearance rates. A case is considered cleared when a law enforcement agency has arrested someone and charged them with the offense, or when the case is cleared by “exceptional means.” Exceptional clearances include situations where the suspect has died, where the victim refuses to cooperate, or where extradition from another jurisdiction is declined.
None of those outcomes require that the case was actually investigated thoroughly. None require that the person arrested was guilty. None require that physical evidence was collected, witnesses were interviewed, or alternative suspects were considered. The clearance rate measures the completion of an administrative step, not the quality of the work that preceded it.
This is not a minor technical distinction. It is the entire measurement problem. A department with a high clearance rate may be arresting the wrong people, closing cases prematurely, or simply operating in a jurisdiction where the nature of local crime makes identification easy. A department with a low clearance rate may be doing rigorous investigative work on complex cases. The metric does not differentiate between these possibilities.
Research has documented that some departments have used “unfounding” — reclassifying reported crimes as not criminal — to manage clearance rate appearances. A crime that is unfounded is removed from the denominator of the clearance rate calculation. The FBI has acknowledged widespread inconsistency in how agencies apply unfounding criteria.
The RAND Study and What It Actually Found
In 1975, the RAND Corporation published what remains one of the most cited studies in criminal justice: “The Criminal Investigation Process,” by Peter Greenwood, Jan Chaiken, and Joan Petersilia. The research examined how detectives across multiple U.S. jurisdictions actually spent their time, and what the relationship was between that work and case outcomes.
The findings disrupted a dominant assumption about investigative work. The study found that the most important factor in whether a case was solved was not detective follow-up. It was whether the victim or a witness at the scene could identify the suspect in the initial report. In the absence of that initial identification, the probability that a case would be solved through subsequent detective work was, in most crime categories, quite low. For property crimes especially, detective follow-up rarely converted an initially unidentified-suspect case into a clearance.
The study also found that detectives spent a substantial portion of their time on cases that had little realistic probability of being solved, and that patrol officers, who take the initial report, were undertrained in the evidence-gathering steps that most directly predict solvability. The bottleneck was at the front end, not the detective stage.
The RAND study’s core finding was that roughly 30% of detectives’ time was spent on cases that could not realistically be solved, and that patrol officers’ initial reports contained the most critical solvability information. Reorganizing how that initial information was gathered would have more impact on clearance outcomes than adding detective capacity.
Subsequent research, including work by John Eck published in 1983 examining the criminal investigation process in Jurisdictions across the country, largely confirmed the RAND findings. Eck’s research for the Police Executive Research Forum found that solvable cases shared predictable characteristics: a named suspect, an identifiable vehicle, traceable property, or specific physical evidence collected at the scene. Cases without these features were rarely solved regardless of detective effort.
Solvability Triage and Who Gets Investigated
The research on solvability criteria revealed something that should have generated more public discussion than it did. Many departments, confronted with more cases than investigators can handle, formally implemented solvability screening. A patrol officer or intake supervisor evaluates the initial report against a checklist of factors. Cases that fall below a threshold score are closed without ever reaching a detective.
The structural problem with this system is not that triage is inherently wrong. Resource allocation decisions are unavoidable. The problem is that solvability criteria encode assumptions about which cases are worth working, and those assumptions have consequences that fall unevenly. Research has documented that cases involving victims from lower-income neighborhoods, cases involving certain crime types, and cases where victims have prior criminal histories receive different treatment under solvability screening than cases with easier initial characteristics.
Eck’s 1983 research noted that the solvability screening systems in place at that time varied substantially across jurisdictions, with some departments using highly structured criteria and others leaving the decision largely to patrol officer discretion. Neither approach was consistently tied to later investigative outcomes. The point was that the screening decision was often invisible to victims and, in many cases, invisible to any external review process.
When a case is closed at the solvability screening stage, the victim typically receives a notification that the case has been closed due to “insufficient leads.” The internal reasoning behind that closure, including what was or was not collected at the initial scene, is not disclosed. There is no standard requirement for external review of solvability closure decisions in most jurisdictions.
Declining Homicide Clearance Rates and What They Reveal
Homicide investigation receives more resources than any other crime category. It is the case type where investigative thoroughness is most consequential and most publicly scrutinized. And it is where the sharpest long-term decline in clearance rates has occurred.
In the 1960s, FBI data reflects a national homicide clearance rate of approximately 90%. By the early 2020s, that figure had declined to approximately 50%, with significant variation by jurisdiction. Research by criminologist Jill Leovy, drawing on Los Angeles homicide records, documented that clearance rates for certain homicide categories, particularly Black male victims in high-crime urban areas, were substantially lower than the aggregate statistics suggest. Leovy’s reporting found that structural underinvestment in certain community types contributed directly to systematically lower clearance rates for those cases.
The homicide clearance decline has multiple proposed explanations in the literature: demographic shifts in urban crime patterns, reductions in community cooperation with police in the wake of documented misconduct and distrust, changes in the nature of homicide from domestic and acquaintance incidents, which have high initial solvability, to gang-related and stranger incidents, which do not. Lawrence Sherman’s research on crime and deterrence has argued that investigative response time and investigative thoroughness on serious violent crimes have measurable deterrence effects. If that relationship holds, declining clearance rates are not just a measurement problem. They are a public safety failure with compounding consequences.
The Documentation Problem
Separate from clearance rates and solvability triage is the question of documentation quality within investigations that do proceed. Research on wrongful convictions provides one lens into this problem.
The National Registry of Exonerations, maintained by the University of Michigan Law School and the Michigan State University College of Law, has catalogued contributing factors in wrongful conviction cases. Among the documented contributors to wrongful convictions: tunnel vision, defined as the premature narrowing of investigative focus on a single suspect before the evidence warrants it; inadequate collection or preservation of physical evidence; failure to document or disclose exculpatory information; and inadequate documentation of witness identification procedures.
These are not rare procedural edge cases. The National Registry data reflects that official misconduct, which includes failures of documentation and disclosure, was a contributing factor in more than half of the exonerations documented as of recent reporting years. That does not mean every case with inadequate documentation results in a wrongful conviction. It means that inadequate documentation is a systemic feature of investigative processes, not an isolated failure.
As of 2024, the National Registry of Exonerations has documented over 3,300 exonerations in the United States since 1989. Research drawing on that registry has identified that perjury or false accusation, official misconduct, and false or misleading forensic evidence are the three most commonly documented contributing factors. All three have a direct relationship to investigative quality and documentation practices.
What the Available Evidence Shows About Investigative Standards
There is no single national dataset that measures what percentage of police investigations meet a defined thoroughness standard. No federal agency audits investigative quality across departments. No uniform documentation requirement exists that would allow such a comparison. What exists instead are several bodies of evidence that function as proxy measures, and each one points in the same direction.
DOJ Pattern-or-Practice Investigations
The Civil Rights Division of the U.S. Department of Justice has authority under 34 U.S.C. § 12601 to investigate law enforcement agencies where there is reason to believe a pattern or practice of civil rights violations exists. These investigations, while not designed specifically to audit investigative thoroughness, have consistently documented systemic investigative failures as part of their findings in major U.S. cities.
The 2016 DOJ investigation of the Baltimore Police Department found that BPD officers routinely failed to document the basis for stops, searches, and arrests, and that this documentation failure was not treated as a disciplinary matter within the department. The investigation found that supervisory review of investigative work was inadequate, that case files frequently lacked required documentation, and that the department had no systematic mechanism for identifying or correcting investigative deficiencies.
The 2017 DOJ investigation of the Chicago Police Department documented that CPD officers were rarely held accountable for using force without justification or documentation, that investigative reports were frequently incomplete or inaccurate, and that the department’s internal accountability mechanisms did not function to identify or correct those failures. The investigation noted that CPD’s systems for reviewing serious incidents produced findings that were not reliable and were not used to drive training or discipline corrections.
The 2015 DOJ investigation of the Ferguson Police Department, while primarily focused on traffic enforcement and municipal court practices, documented that investigative practices involving arrest and charging were frequently not grounded in adequate evidentiary review, and that officers routinely applied charges without the documentation that would support them at trial.
These are not marginal departments or unusual findings. Baltimore, Chicago, and Ferguson were each selected for investigation because the documented problems were visible and persistent. The DOJ investigations describe investigative and documentation failures not as individual officer mistakes but as systemic, department-level practices that had continued across years and administrations.
DOJ pattern-or-practice investigations are not random audits. They are triggered by sustained evidence of systemic failure. The fact that investigations of Baltimore, Chicago, Cleveland, Ferguson, New Orleans, and Seattle, among others, all documented investigative and documentation failures as core findings suggests these are not outliers. They are the visible portion of a broader structural problem.
The Sexual Assault Kit Backlog as a Proxy Metric
One of the most concrete measures of investigative failure at scale is the sexual assault kit backlog. A rape kit that is collected but never submitted for testing represents a forensic investigation that was not conducted, for a crime that was reported, against a victim who cooperated with the process. The kit sits in a storage facility. The case sits open or is closed for insufficient leads.
In 2019, the Bureau of Justice Statistics published findings from a national census of sexual assault kits held in law enforcement agencies and crime labs. The census found that approximately 397,000 sexual assault kits were being held untested in law enforcement agencies and crime labs as of that date. That figure represented kits collected from reported crimes that had not been submitted for forensic analysis. It did not include kits that had been submitted but were awaiting processing in crime lab backlogs.
The backlog is a documentation of a decision, made at some point in each of those cases, that forensic investigation would not proceed. In some cases that decision reflected a reasonable prosecutorial assessment that the case could not proceed. In a substantial number of cases, research by the Joyful Heart Foundation and the End the Backlog initiative documented that kits were never tested simply because the department lacked the resources, the protocols, or the internal pressure to submit them.
Detroit’s experience with its rape kit backlog is instructive and has been publicly documented. In 2009, a warehouse containing more than 11,000 untested rape kits was discovered. Michigan received federal funding to begin testing those kits. By 2020, testing had identified more than 800 serial rapists, including perpetrators linked to crimes in multiple states. The kits had sat untested for years, in some cases for decades. Each one was an investigation that did not happen.
Testing of Detroit’s 11,000-plus backlogged rape kits identified more than 800 suspected serial rapists, with DNA matches linking perpetrators to crimes in at least 40 states and Washington D.C. These were not cold cases where the evidence was unavailable. They were cases where the evidence existed and was not used. The backlog was an investigative failure that had been warehoused.
Homicide Detective Caseloads Against Research Standards
The Police Executive Research Forum has published guidance on homicide investigation recommending that homicide detectives carry no more than five to eight open cases at one time to allow for thorough investigation. This recommendation is grounded in research on what investigative tasks are required for a thorough homicide investigation and how long those tasks take when done properly.
Surveys of major urban departments have documented that many operate well above that threshold. A 2019 report by the nonprofit Murder Accountability Project found that in numerous large U.S. cities, detectives were carrying caseloads two to four times the recommended maximum. Research by the University of Pennsylvania’s Jerry Lee Center for Criminology has documented the relationship between investigative caseload and case outcomes, finding that caseload pressure is associated with shorter investigation timelines and lower clearance rates.
Caseload is a throughput constraint. A detective carrying 20 open homicide cases cannot spend the same time on each case that a detective carrying six can. The investigative steps that are most time-intensive, including canvassing for witnesses, reviewing surveillance footage, pursuing forensic leads, and re-interviewing witnesses as new information develops, are the steps most likely to be compressed or skipped when caseload pressure is high. This is not a judgment about individual detective effort. It is a structural constraint that produces predictable investigative shortfalls at scale.
The Exoneration Record as an Investigative Audit
The National Registry of Exonerations provides the most granular available evidence of what investigative failure looks like at the individual case level. The Registry documents contributing factors for each exoneration. As of its most recent annual report, official misconduct was identified as a contributing factor in approximately 54% of exoneration cases. Perjury or false accusation was a factor in approximately 56%. False or misleading forensic evidence was a factor in approximately 24%.
These are not three separate categories of failure. They frequently co-occur and are often downstream of the same investigative decisions made early in a case. A detective who prematurely narrows focus to a single suspect generates pressure to build an evidentiary record that supports that conclusion. That pressure can produce or tolerate witness coaching, suppress contradictory evidence, and accept forensic findings without adequate scrutiny. The Registry’s data reflects the outcomes of that process across more than three decades of documented cases.
The exoneration record is not a representative sample of all investigations. It is a documented record of the worst investigative failures, the ones severe enough to produce a conviction that was later reversed. It cannot tell us what percentage of investigations are thorough. It can tell us that thorough investigation is not the baseline the system produces in every case, and that the structural conditions that produce investigative failure, caseload pressure, inadequate documentation standards, and absent external review, are not aberrations.
What Thorough Investigation Actually Requires
The research literature has been clearer about what predicts investigative failure than about what constitutes investigative thoroughness. But several consistent structural requirements emerge from the evidence base.
The RAND study and subsequent research consistently find that the quality of information gathered at the initial scene is the single largest predictor of solvability. Patrol officer training in evidence collection, witness identification, and initial report documentation has more impact on investigative outcomes than detective caseload or specialization alone.
Research by the Police Executive Research Forum has found that detective caseloads in many jurisdictions far exceed what allows for thorough investigation. PERF has noted that recommended standards for homicide detective caseloads suggest no more than five to eight open cases per detective at one time. Many urban departments operate well above that threshold.
Case file documentation standards vary by jurisdiction and are rarely subject to external audit. Research on wrongful convictions and civil rights litigation has documented that the absence of documentation standards creates conditions where investigative decisions cannot be reconstructed or reviewed, either internally or by defense counsel.
Eck’s research and subsequent work by the Police Executive Research Forum has noted that solvability screening decisions are rarely subject to supervisor review or documentation requirements. A decision to close a case without detective follow-up can be made by a single patrol officer or supervisor without any recorded justification or appeal mechanism for the victim.
The Counterargument
The research on investigative thoroughness is sometimes countered with the argument that clearance rate declines reflect changes in crime patterns, not investigative failures. There is real substance to this. The shift toward stranger-involved violent crime, the documented decline in community cooperation with law enforcement in jurisdictions where distrust is high, and resource constraints on investigative units are all legitimate contributing factors that are not reducible to investigative failure.
A second counterargument notes that the RAND study is now more than 50 years old, and that investigative technology, including DNA evidence, digital forensics, and surveillance infrastructure, has changed what investigators can do with limited initial information. That is also partially true. DNA evidence has been directly responsible for solving cold cases and has contributed to both convictions and exonerations. Digital forensics has expanded the evidentiary record available to investigators.
Neither argument, however, addresses the core problem identified in the literature: that clearance rates measure administrative resolution rather than investigative quality, that documentation standards are inconsistent and rarely externally audited, that solvability screening decisions are invisible to victims and to review, and that the resources allocated to investigation are, in many jurisdictions, insufficient for thorough case work at any technology level. The counterarguments improve at the margins. They do not resolve the structural problem.
Why This Case Matters
The gap between how police investigations are publicly represented and how they actually function is not a marginal policy concern. It is the mechanism by which the wrong person gets arrested, by which the right person goes unidentified, and by which communities in which crime is concentrated and investigative resources are thin receive systematically worse outcomes from the criminal justice system.
The research on this question is substantial, consistent, and several decades old. The policy response to it has been partial at best. Clearance rates remain the dominant public metric. Solvability screening decisions remain largely invisible. Detective caseloads in many jurisdictions remain far above what the research recommends. Documentation standards remain inconsistent and rarely audited from outside the department.
Thorough investigation is not an aspirational standard. It is the baseline the system claims to operate at. The evidence says otherwise.
Sources
Bluebook (Legal) Williams, Rita, How Thorough Are Police Investigations, Really?, Clutch Justice (June 4, 2026), https://clutchjustice.com/.
APA 7 Williams, R. (2026, June 4). How thorough are police investigations, really? Clutch Justice. https://clutchjustice.com/
MLA 9 Williams, Rita. “How Thorough Are Police Investigations, Really?” Clutch Justice, 4 June 2026, clutchjustice.com/.
Chicago Williams, Rita. “How Thorough Are Police Investigations, Really?” Clutch Justice, June 4, 2026. https://clutchjustice.com/.
If you have documents and a situation that doesn’t add up, a forensic record review maps the contradictions, identifies the gaps, and produces a written findings memo you can act on — in 24 hours or less.
