Workplace Violence in Cannabis Retail: Reading the Public Data

Cannabis retail safety conversations often begin with a robbery story or a viral video. The story is memorable, but it is not a risk assessment. A dispensary needs to know what kinds of events are plausible at its location, which controls reduce exposure, and what evidence shows the controls are working. Public data can inform that work, but only if the operator understands what the data measures.

This review examines three public sources: the Bureau of Labor Statistics Census of Fatal Occupational Injuries, the FBI's Uniform Crime Reporting program, and OSHA's workplace violence guidance. It explains the useful signals, the coverage gaps, and the questions a dispensary should answer with its own records. The conclusion is not that public data is useless. It is that public data is a starting frame, not a local diagnosis.

The measurement problem

The first problem is classification. Public occupational datasets group employers by industry codes, and cannabis retail does not always have a clean, separate category. A dispensary may be classified within retail trade, health and personal care, or another category depending on the state, the business license, and the reporting system. A fatality or injury at a dispensary may be included in a broader retail count without a cannabis-specific label.

The second problem is underreporting. Workplace violence is not always reported to police, and a near miss is rarely reported at all. A budtender who is threatened by a customer may not file a police report, especially if the customer leaves and the store wants to avoid attention. The public crime data may show a robbery but not the lower-level aggression that precedes it.

The third problem is denominator. A national rate per 100,000 workers is not a prediction for a single store. The store's hours, layout, cash handling, location, staffing, and customer mix matter more than the national average. A rate can help an operator ask better questions, but it cannot set the store's security budget by itself.

What the BLS data shows

The BLS Census of Fatal Occupational Injuries summary for 2024 reported 5,070 fatal work injuries in the United States, down 4.0 percent from 5,283 in 2023. The fatal injury rate was 3.3 fatalities per 100,000 full-time equivalent workers, down from 3.5 in 2023. The release also reported that a worker died every 104 minutes from a work-related injury.

Those headline numbers are a national baseline. The more relevant detail for retail safety is the violent act category. BLS reported 733 fatal violent acts in 2024, down from 740 in 2023. Homicides accounted for 64.1 percent of that category, with 470 fatalities, up from 458 in 2023. The release also noted that women accounted for 8.1 percent of all worker fatalities but 15.3 percent of fatalities due to homicides.

The BLS tables allow a more granular look by occupation, industry, and event. The fatal injury tables are the right place to check the current retail categories and the definitions behind the counts. An operator should read the table notes, not just the headline. A category that includes "retail trade" may combine gas stations, grocery stores, and specialty retail in ways that matter for interpretation.

What the BLS data does not show is the cannabis-specific experience. It cannot tell a dispensary whether its own late-night shift is more or less exposed than a convenience store, a liquor store, or a pharmacy. It can show that homicide is a real category of work-related death and that retail customer-facing work deserves attention. It cannot assign a probability to a specific address.

What the FBI data adds

The FBI's Uniform Crime Reporting program collects crime data from participating agencies. The Crime Data Explorer allows users to examine reported offenses by location and offense type. For a dispensary, the relevant categories may include robbery, aggravated assault, burglary, and larceny.

The value of the FBI data is local comparison. An operator can look at reported robbery and assault patterns in the jurisdiction and compare them with the store's incident log. If the store reports more customer aggression than the surrounding retail area, the difference may point to store-specific factors: layout, hours, cash visibility, queue management, or the way staff respond to conflict.

The limitation is that police data reflects reports, not events. A store with a strong reporting culture may look worse than a store where incidents go unreported. A jurisdiction may have incomplete participation in UCR or may change its reporting system during the period. An operator should treat the data as one input and document the source, the query date, and the geography.

What OSHA guidance contributes

OSHA's workplace violence guidance describes a prevention program built on management commitment, worker participation, hazard identification, hazard prevention and control, training, and evaluation. The framework is useful because it is not limited to robbery. It covers threats, verbal abuse, physical assaults, and other violent acts that may occur in a customer-facing setting.

For a dispensary, the hazard identification step should include a walk-through of the customer journey: entrance, queue, ID check, sales floor, checkout, exit, delivery handoff, and closing. At each point, the team should ask what could go wrong and what control exists. A camera may help after an event, but it does not prevent the event. A well-placed greeter, a clear queue, a cash-handling policy, and a trained de-escalation response may matter more.

OSHA's guidance is advisory for many employers, and the applicable legal duties depend on the jurisdiction and the facts. The operator should confirm the requirements with counsel. The practical value is the program structure: identify, control, train, and review.

Building a local risk picture

A dispensary cannot rely on national data to decide whether to add a security guard, change closing hours, or install a new camera. It needs a local risk picture built from four sources:

• Incident reports, including threats and near misses
• Police reports and calls for service, if available
• Staff interviews and exit feedback
• Physical security and cash-handling observations

The incident report should capture the time, location, people involved, what happened, what the staff did, whether anyone was injured, whether police were called, and what changed afterward. A near-miss field is essential. A customer who pounds on the counter and leaves is a near miss, not a non-event.

The security incident handoff record can provide the intake structure. The front-of-house support team can maintain the log and the follow-up, but the store manager and security owner should review the pattern.

Leading indicators to watch

A local review should look for changes in the conditions that precede violence, not just the count of violent events. Useful indicators include:

• Increase in after-dark customer aggression reports
• Repeated conflicts at the ID-check or payment step
• Long queues without a visible staff presence
• Cash handling that is visible from the entrance
• Delivery handoffs at unlit or uncontrolled locations
• Staff turnover in closing shifts
• Repeat customers who have been warned about conduct

None of these indicators proves that violence will occur. They are signals that a control may be weakening. The team should record the observation, assign an owner, and review the response at the next safety meeting.

A small study design

A dispensary that wants a more rigorous local picture can run a 90-day review. Define the eligible events in advance: physical assaults, threats, aggressive gestures, harassment, robbery, attempted robbery, and near misses. Use one intake form and one location. Report counts, rates per 1,000 customer visits if the POS can provide a denominator, and the percentage of events with a documented response.

Do not compare the store to a national rate as if the populations were identical. Compare the store to its own prior period, and describe any change in reporting or staffing. If the store adds a new incident form, the count may rise because reporting improved, not because the store became more dangerous. That is a measurement issue, not a failure.

Limitations and responsible use

This review uses public data to frame questions. It does not establish that cannabis retail is more or less dangerous than another retail category, and it does not predict what will happen at a specific store. The BLS categories are broad, the FBI data depends on reporting, and the OSHA guidance is a framework rather than a local risk score.

The results should not be used to stigmatize customers, neighborhoods, or workers. A safety program should reduce harm without turning the store into a hostile environment. The operator should involve staff in the design, protect their privacy, and avoid using incident reports as a disciplinary tool for people who report honestly.

Practical conclusion

Public data is most useful when it prompts a local question. The BLS homicide count asks whether the store has a violence prevention program. The FBI robbery data asks whether the store's cash and closing routines are visible to the street. The OSHA framework asks whether the team has trained, documented controls and a review cycle.

The dispensary's own record is the missing layer. Start with a 90-day incident and near-miss log, review it with the people who work the floor, and fix the conditions that show up repeatedly. That is a more defensible safety strategy than a national statistic applied to a local address.

Sources

1. BLS Census of Fatal Occupational Injuries Summary, 2024
2. BLS fatal injury tables
3. BLS Injuries, Illnesses, and Fatalities overview
4. FBI Uniform Crime Reporting program
5. FBI Crime Data Explorer
6. OSHA workplace violence guidance
7. CISA physical security resources

Internal reading: security incident handoff and front-of-house support.