Introduction and Research Purpose
This research uses secondary data to examine how reported police brutality, excessive force, and serious use-of-force incidents have changed over time. Secondary data are records originally collected by governments, researchers, courts, oversight bodies, journalists, or civil-rights organizations for purposes other than this particular paper. They include police-public contact surveys, use-of-force reporting systems, mortality records, court decisions, consent decrees, inspector-general findings, peer-reviewed studies, policy documents, and carefully documented incident databases. This approach is appropriate because the research question is historical and comparative. Past encounters cannot be recreated for direct observation, and a single researcher could not interview thousands of officers, residents, departments, courts, hospitals, and families across multiple jurisdictions within a normal academic schedule.
The study must define its central concepts carefully. “Police brutality” is a broad moral and political expression commonly used for cruel, unnecessary, discriminatory, or excessive conduct by law-enforcement personnel, while “excessive force” is often a legal term evaluated under the law of a particular jurisdiction. A complaint, injury, firearm discharge, or death does not by itself establish that a court would find force unlawful. Conversely, the absence of a criminal conviction does not prove that conduct was necessary, ethical, or consistent with departmental policy. For this reason, incidents should first be coded through observable facts—type of force, injury, circumstances, available evidence, and subsequent administrative or judicial findings—before they are placed into broader evaluative categories.
Why Secondary Data Are Necessary
Secondary evidence allows repeated measurements across time, which is essential when the purpose is to assess change rather than describe one moment. Researchers can compare periods before and after reporting reforms, body-camera adoption, consent decrees, policy changes, major public controversies, or shifts in training and supervision. The method can also help distinguish changes in visibility from changes in underlying conduct. An increase in recorded incidents after mandatory reporting or expanded video coverage may indicate better documentation rather than a true increase in force, while an apparent decline can reflect changes in public contact, classification, or agency participation. Historical interpretation therefore requires information about the data system itself as well as the incidents it records.
Official national sources provide useful scale but have important limitations. The FBI’s National Use-of-Force Data Collection was created to develop national information about incidents involving death, serious bodily injury, or firearm discharge at or toward a person. In its August 2025 update, the FBI reported that participating agencies covered about 78% of the law-enforcement population, below the 80% coverage threshold used for releasing national incident counts, so incomplete participation remained an important limitation. The Bureau of Justice Statistics Police-Public Contact Survey provides a different perspective because it asks a nationally representative sample of U.S. residents about contact with police, including perceived misconduct, threats of force, and nonfatal force. The 2022 survey, published in 2024, offers a denominator based on public contact rather than only reported agency incidents (Federal Bureau of Investigation, 2025; Bureau of Justice Statistics, 2024).
Mortality systems, departmental reports, court files, and Department of Justice investigations add other forms of evidence. Public-health records can identify fatal outcomes, although a “legal intervention” classification does not by itself determine whether force was legally justified. Local departments may publish more detailed information about tactics, injuries, complaints, and disciplinary outcomes, but definitions can change over time and agencies may count events differently. Department of Justice pattern-or-practice investigations can provide detailed case studies of supervision, training, reporting, discipline, crisis response, and discrimination where federal investigators found systemic constitutional problems. Such departments are not a random sample of all agencies, so their findings should be used to understand mechanisms rather than generalized automatically to every jurisdiction.
Longitudinal and Comparative Research Design
A defensible study should define a period long enough to observe change but narrow enough to maintain comparable data. A ten- or fifteen-year United States study may be more manageable than an unrestricted global comparison because policing structures, firearm prevalence, legal standards, data transparency, and definitions of force differ sharply across countries. If cross-national examples are included, each jurisdiction should be described within its own legal and institutional context rather than ranked through measures that are not equivalent. The unit of analysis also needs to remain consistent. An incident, agency-year, jurisdiction-year, court finding, or reform case answers a different question, and mixing these units without explanation can create misleading conclusions.
Comparison should account for opportunity and exposure. Raw incident totals can be misleading because larger cities may report more uses of force simply because they have more residents, officers, arrests, or police-public contacts. Population rates show community burden, while contact-based or arrest-based rates address risk during particular types of interaction. Each denominator has limitations. Arrest-based rates, for example, can reproduce differences created by enforcement patterns. Where data permit, the study should present more than one denominator and explain what each rate means rather than treating one calculation as a complete measure of fairness or necessity.
Policy and organizational comparisons can examine whether changes in supervision, reporting, de-escalation requirements, body-camera rules, duty-to-intervene provisions, early-intervention systems, and disciplinary procedures coincide with changes in outcomes. Such analysis should not assume causation simply because a reform precedes a decline. Reforms are often adopted in response to existing problems, and implementation may vary significantly. A written policy can have little effect when training, staffing, supervision, data quality, or enforcement remain weak. The research therefore needs both quantitative trends and qualitative evidence explaining how policies were actually applied.
Coding, Analysis, and Source Appraisal
Each incident should be coded with variables that can be observed or verified: date, jurisdiction, reason for encounter, type and level of force, injury, fatality, reported threat, weapon information, resistance, mental-health indicators where documented, number of officers and civilians, available video, demographic information, complaint status, administrative findings, civil outcomes, criminal outcomes, and relevant policy changes. Missing information should remain marked as missing rather than guessed. Resistance and threat may be relevant to evaluating proportionality, but they should not be used to suggest automatically that civilians cause officers to use excessive force. Officers retain professional and legal responsibilities to use proportionate force, provide required medical assistance, document their actions accurately, and stop using force when the justification ends.
Every source should also be evaluated for authority, purpose, definition, geographic coverage, completeness, update frequency, and possible bias. Official status does not guarantee complete reporting, and advocacy status does not make a source unreliable. The central question is whether the source explains its method and supports the specific claim being made. Departmental accounts can be compared with court records, medical findings, video, independent investigations, and community reporting. Journalism can identify incidents and patterns before official datasets are updated, but articles should be checked against primary records where possible. Independent databases may improve visibility, especially for fatal encounters, while still requiring review of inclusion rules, verification procedures, duplicate handling, and historical coverage.
Statistical analysis should distinguish descriptive patterns from causal conclusions. Annual rates can be graphed or compared across policy periods, while sensitivity analyses can test whether conclusions change when agencies with incomplete reporting are excluded. Demographic disparities should be reported with relevant denominators and context rather than treated automatically as proof of individual intent. Statistical controls can help examine differences in exposure, enforcement practices, neighborhood assignment, and encounter characteristics, but they do not make unequal outcomes morally neutral or prove that all relevant factors have been measured. The aim is to identify patterns accurately while keeping the limits of inference visible.
Limitations, Ethics, and Interpretation
The central limitation of this research is inconsistent and incomplete data. Voluntary reporting, local definition changes, sealed records, underreporting by civilians, incomplete demographic information, and stronger documentation of fatal incidents than nonfatal encounters make precise trend claims difficult. The FBI’s 2025 update illustrates this problem because national reporting still had not reached the participation threshold needed for complete incident counts. Survey data solve some problems by capturing experiences that never become administrative complaints, but they introduce others such as recall, sampling limitations for rare events, and exclusion of fatal encounters. No single dataset can answer every question about police force.
Secondary sources also cannot capture every perception and decision made during a rapidly unfolding encounter. Officer reports may emphasize threat, civilian accounts may emphasize harm, and video may omit events outside the frame. Triangulation improves confidence but may not resolve every disagreement. The final paper should distinguish confirmed facts, disputed claims, legal findings, policy findings, and the researcher’s interpretation. It should also avoid language that assumes guilt before evidence or treats institutional criticism as hostility toward every officer. Neutral methodology is especially important in a politically contested subject because the purpose of research is to clarify evidence rather than predetermine a conclusion.
Ethical presentation matters as much as statistical technique. Victims, officers, families, and communities should not be reduced to sensational examples. Graphic material should be used only when analytically necessary, and personal information should be limited to what is public and relevant. Secondary analysis reduces the need to expose participants to difficult interviews about injury, death, discrimination, or trauma, but it is not ethically risk-free. Incident-level records can contain names, addresses, medical details, and video. Researchers should follow privacy, licensing, and institutional-review requirements and explain uncertainty honestly.
Conclusion
Secondary data are appropriate for studying the evolution of police brutality and excessive force because the question spans multiple years, agencies, data systems, and legal settings. Official use-of-force reports, police-public contact surveys, mortality records, local departmental data, court files, Department of Justice investigations, peer-reviewed research, journalism, and independent databases each contribute different kinds of evidence. Their value depends on clear definitions, consistent units of analysis, appropriate denominators, careful source appraisal, and explicit treatment of missing information.
A strong study should not assume in advance that brutality has universally increased or decreased. It should separate changes in documentation from changes in underlying conduct, distinguish reported force from legal or administrative findings, and examine how supervision, policy, training, reporting, discipline, and community conditions relate to variation across jurisdictions. The most credible conclusion will be the one supported by converging evidence while remaining transparent about what the available data cannot establish.
References
Bureau of Justice Statistics. (2024). Contacts Between Police and the Public, 2022. U.S. Department of Justice.
Federal Bureau of Investigation. (2025). National Use-of-Force Data Collection Update, June 2025.
National Academies of Sciences, Engineering, and Medicine. (2018). Proactive Policing: Effects on Crime and Communities. National Academies Press.
U.S. Department of Justice, Civil Rights Division. Conduct of Law Enforcement Agencies: Pattern-or-Practice Investigations.
Cite This Work
To export a reference to this article please select a referencing stye below:
Academic Master Education Team is a group of academic editors and subject specialists responsible for producing structured, research-backed essays across multiple disciplines. Each article is developed following Academic Master’s Editorial Policy and supported by credible academic references. The team ensures clarity, citation accuracy, and adherence to ethical academic writing standards
Content reviewed under Academic Master Editorial Policy.
