March 28, 2026
How Definitions, Systems, and Reporting Methods Shape the Numbers
A Common Question
As new reports on veteran suicide are released at the national, state, and local levels, a familiar question follows: Why don’t the numbers match?
The differences can create confusion and, at times, concern about whether the data can be trusted at all. The reality is more straightforward. The issue is not that the data is wrong. The issue is that the systems producing the data are different by design.
These differences are not the result of error or manipulation. They reflect how data is collected and classified across systems. Even small variations in how veteran status is recorded, or how location is assigned, can meaningfully change the results.
Different Measures
At first glance, it can seem as if different systems are measuring the same thing but arriving at different answers. In reality, they are not measuring the same thing in the same way. They start from different assumptions, use different definitions, and are built for different purposes.
The starting point matters. County-level data, drawn directly from death certificates, is closest to the original record. However, even the accuracy of information captured on death certificates – a primary information source – depends on the information available at the time. Sometimes, the grieving next of kin may assume their elder relative was a veteran, or in some cases, if the veteran kept their service private, that detail may be missed. Other demographics, such as occupation, can be misreported.
Several structural factors shape how the numbers are ultimately reported:
Definition of a Veteran
The federal veteran definition is based on federal active duty service and status at the time of death. State definitions rely on what is recorded on a death certificate, including what families report.
Geography
Another area of difference is how the location is recorded. Some systems count deaths based on where a person lived, while others count them based on where the death occurred.
Data Flow
Information typically flows from county to state to federal systems. At each step, data is standardized for comparison across jurisdictions, which can result in a loss of detail and precision.
Cause of Death
Another factor that shapes the data is who is responsible for determining the cause and manner of death. Across the country, this role is handled through two different systems: medical examiners and coroners. The distinction matters because it influences how a death is ultimately classified, including whether it is determined to be suicide, natural, accidental, or undetermined.
Aspect | Medical Examiner System | Coroner System |
Role structure | Led by appointed officials | Led by elected officials |
Primary background | Physicians (often forensic pathologists) | Varies by jurisdiction; not always required to have a medical background |
Basis for determination | Medical evaluation of history, injuries, and toxicology | May rely on available evidence, with medical input as needed |
Use of medical expertise | Built into the role and process | Often accessed through consultants or external medical professionals |
Consistency of approach | More standardized within jurisdictions using this model | Can vary by county or state, depending on structure and resources |
Variation across locations | Differences still exist, but generally within a medical framework | Greater variability in structure, qualifications, and processes across jurisdictions |
How this affects classification | Determinations are made within a medically driven process | Determinations are made within locally defined processes that may differ across jurisdictions |
Arizona's System
In Arizona, most deaths are investigated through county-based medical examiner systems, particularly in the state’s largest counties. In this model, death investigations are led by physicians, typically forensic pathologists, who assess medical history, injury patterns, toxicology, and other available evidence when determining cause and manner of death.
Why Classification Differs Across the Country
In other parts of the country, a coroner system is used. In those systems, the coroner is often an elected official. Requirements for the role vary by jurisdiction, and coroners may work with medical professionals as part of the investigation process.
Another factor that contributes to variation in veteran suicide data is that there is no single, standardized national system for how deaths are investigated and classified.
Definitions, procedures, and thresholds for determining cause and manner of death can vary across states and even across counties. That variation is not just about whether a jurisdiction uses a medical examiner or coroner system. It is also shaped by the resources available, access to forensic pathology and toxicology, staff training, case volume, and local protocols. In practice, those factors often matter more than the system label. When data is brought together from across the country, variation in how deaths are classified should be expected.
Data Moves Through Systems
Now, let’s think about the data as it moves through the system. Information typically begins at the local level, often at the county, where death certificates are completed. From there, it is aggregated at the state level and eventually incorporated into national datasets. At each step, the data is standardized so it can be compared across all states and territories.
That standardization serves a purpose. It allows for national trend analysis and policy development. But it comes with a tradeoff. As data is standardized, some of the detail that exists at the local level is lost. Cases may be categorized differently. Some information may not translate cleanly between systems. In some instances, key fields, such as veteran status, location, or even occupation, may be incomplete or inconsistently captured.
It is important to remember that this standardization is not a flaw in any one system. It is a function of how complex, multi-layered data systems operate. Put simply, the further the data gets from its original source, the less precise it becomes. That does not make the data unusable. It means it must be interpreted within the context of how it was created.
Reading the Data in Context
Differences in data should not undermine confidence in the work. Rather, understanding that the differences are there and reflect how the various systems function can lead to a clearer understanding of how the data are constructed and confidence that differences are not necessarily errors. Once you understand how the data is built, the next step is knowing how to read it.
Do
- Expect variation across systems
- Look at patterns and trends over time
- Ask what question the data is answering
- Use local and state data when context is needed
- Understand how the data was defined and collected
Avoid
- Avoid assuming all sources are measuring the same thing
- Avoid equating national data with state/local data
- Avoid expecting federal data to exactly match state data
- Avoid focusing on small differences between reports
- Avoid assuming one dataset is “the correct number”
Different Systems Produce Differences in Data
Local data is often closest to the original record and can provide more detail. National data allows for comparison across states but requires standardization that can reduce that level of detail. Both serve a purpose, but they should not be expected to produce identical results.
These numbers aren’t competing with each other. They’re built in different systems, using different definitions, and shaped by different processes. When they don’t match, that doesn’t mean something is wrong. They’re reflecting different parts of the same reality.
Clarity Supports Better Understanding
As this work continues, clearly outlining definitions, assumptions, and methodology as part of published reports can help ensure the data is understood as intended. When report authors include that context up front, it becomes easier to see how the numbers were built, what they represent, and how they should be used. That level of transparency helps reduce confusion and allows different data sources to be interpreted within the right context.
Recap
- The numbers are not measuring the same thing in the same way
- Differences in definitions and classification change who is counted
- Local data is closer to the original record, but not perfect
- National data is more standardized, but less precise
- When the numbers don’t match, it doesn’t mean something is wrong
