8, 10, or 12 hours: what the research actually says about police shift patterns.
Departments spend months arguing about shift length. A federally funded randomized trial has already measured most of what gets argued about — and the results are more surprising, and more useful, than either side of that argument usually expects.
About this page. We build scheduling software. We have never run a patrol division, and nothing here is advice on how to police your jurisdiction.
What follows is the published research with citations so you can read the originals, plus the scheduling and overtime arithmetic we do know well. Where the two meet is your judgment, not ours.
On most of what gets argued about, shift length made no measurable difference.
The Police Foundation, now the National Policing Institute, ran a randomized controlled experiment funded by the National Institute of Justice. Researchers randomly assigned 275 officers in Detroit, Michigan and Arlington, Texas to work five 8-hour days, four 10-hour days, or three 12-hour days for six months, across day, evening, and midnight shifts.
Across work performance, health, and work-family conflict, the study found no significant differences among the three shift lengths. Reported sleep quality and reported sleep disorders did not differ either.
That is worth sitting with before a department commits a year to the debate. The outcomes most often used to argue for or against a rotation are the outcomes the experiment could not distinguish.
Three things did differ, and they are the subject of the rest of this page: how much sleep officers got, how alert they reported being, and how much overtime they worked.
What differed
Sleep amount. Quality of work life. Reported sleepiness and alertness. Overtime hours worked.
What did not
Work performance. Safety measures including driving and reaction time. Health. Work-family conflict. Sleep quality. Reported sleep disorders.
Four rotations, compared on what a scheduler has to plan around.
Crew counts below follow the standard published configurations for each rotation. Hours are the pattern’s own arithmetic, before any overtime, holdover, or callback.
| 5 on / 2 off, 8 hr | 4 on / 3 off, 10 hr | Pitman 2-2-3, 12 hr | 4 on / 4 off, 12 hr | |
|---|---|---|---|---|
| Cycle length | 7 days | 7 days, or 20 days in the 4-2 4-3 4-3 variant | 14 days | 8 days |
| Crews typically used | 3 watches plus relief | 5 or 6 teams with overlapping shifts | 4 crews | 4 crews |
| Average hours per week | 40 | 40 | 42 | 42 |
| Days worked per year | 260 | 208 | 182.5 | 182.5 |
| Longest run of work days | 5 | 4 | 3 | 4 |
| Shift changes per day | 3 | 3, with built-in overlap | 2 | 2 |
| Weekend structure | Depends on the build | Depends on the build | Every other weekend fully off | Rotates through the week |
| Overtime worked in the experiment | Highest of the three | Lowest of the three | Between the two, closer to the 10-hour result | |
| Reported alertness | Reference group | No significant difference from 8-hour | Significantly lower than 8-hour; more reported sleepiness | |
Scroll the table sideways to compare all four rotations.
Overtime and alertness rows report findings from Amendola et al. and describe 8-, 10-, and 12-hour shift lengths as studied, not these four specific rotations. The experiment tested shift length, not named patterns.
The shortest shift produced the most overtime.
This is the result most likely to matter to a budget, and the one most often reported incorrectly.
A note on a figure you may see reversed elsewhere
These two multipliers are frequently published the other way around, as though 8-hour officers worked three times the overtime of the 10-hour group and five times that of the 12-hour group. Both versions circulate, sometimes within a single article.
The National Institute of Justice, which funded the study, reports that officers on 8-hour shifts worked more than five times as much overtime as those on 10-hour shifts and more than three times as much as those on 12-hour shifts. That ordering is the one consistent with the study’s other published figure — that the 10-hour group worked about 80 percent less overtime than the 8-hour group, which is a factor of five, not three.
We have used the NIJ ordering throughout. Both sources are linked at the foot of this page so you can check the arithmetic yourself rather than take our word for it.
The mechanism the researchers point to is not mysterious. Shorter shifts mean more shift changes, and every changeover is another opportunity for an officer to stay late finishing a call or paperwork. Longer shifts with built-in overlap absorb that work inside the scheduled day instead of paying for it afterward.
The 10-hour group slept more. The 12-hour group reported being less alert.
Officers working 10-hour shifts got significantly more sleep per night than those on 8-hour shifts — more than half an hour. The lead researcher put the annual effect at over 150 hours of additional sleep; the National Policing Institute has since described it as nearly 185 hours.
Officers on 12-hour shifts reported greater sleepiness and lower alertness at work than the 8-hour group. The researchers noted that people tend to underestimate their own fatigue, and that accident risk rises with hours worked, and advised caution when considering 12-hour shifts.
Quality of work life followed the same shape: lowest among 8-hour officers, highest among 10-hour officers, with no measurable benefit from 12-hour shifts.
Why alertness is not the same as performance
The lower alertness reported by 12-hour officers did not translate into worse measured performance or more negative outcomes in this study. Driving and reaction-time measures did not differ.
That distinction matters when the finding gets quoted in a labor-management meeting. What the experiment found was a difference in how alert officers reported feeling, not a demonstrated drop in how well they did the job.
The base rotation is rarely what makes an officer tired.
A pattern is a plan. What officers actually work is that plan plus holdovers, callbacks, mandatory overtime, court, and off-duty employment. The published fatigue research concerns the total, not the rotation on paper.
The impairment benchmark
Being awake for 17 hours impairs performance to a degree comparable with a blood alcohol concentration of 0.05 percent. At 24 hours the comparison is 0.10 percent, beyond the legal limit for driving.
That benchmark is what turns a 16-hour day from a scheduling inconvenience into a measurable question.
How far the total can drift
A published account of one department’s overtime found that its highest overtime earners regularly logged 16-hour days, with a single officer working a 16-hour shift at least 68 times across a five-year period.
No rotation produces that. Accumulation does — and it concentrates on a small number of people.
Fatigue and decision-making
Laboratory simulations of officer-citizen encounters found that tired officers were more likely to use force or escalate, and less likely to employ de-escalation tactics, than when they were rested.
Separately, officers working night duty were found to have twice the rate of long-term injury leave of 90 days or more compared with afternoon-shift officers, and three times the rate of day-shift officers.
The prevalence problem underneath all of it
A separate study screened 4,957 state and local law enforcement officers across the United States and Canada. Just over 40 percent had at least one sleep disorder, most of which had never been diagnosed — obstructive sleep apnea in 33.6 percent, moderate to severe insomnia in 6.5 percent, and shift work disorder in 5.4 percent. Sleep disorders are roughly twice as prevalent among police officers as in the general public.
Nearly 46 percent of all participants reported having fallen asleep while driving, and about a quarter said it happened one to two times a month. Officers with a sleep disorder were more likely to make serious administrative errors or safety violations, and to report uncontrolled anger toward suspects.
What the National Policing Institute recommends.
These are their recommendations, not ours. We reproduce them because they are the clearest published summary of what the evidence supports, and because several of them are decisions a scheduling system has to be told about.
On the rotation itself
Consider 10-hour shifts, four on and three off, as a strong option for patrol. Where 12-hour shifts are used, implement them cautiously, keep them voluntary, and monitor for signs of fatigue. Avoid quick shift rotations, and where rotation is necessary use forward rotation — day to evening to night — which is better tolerated than backward rotation.
On limits
Set maximum daily and weekly hours. Many experts recommend no more than 12 hours in a day and 60 hours in a week as an upper bound. Require a minimum rest period between shifts, in the range of 8 to 10 hours. Count off-duty employment against those limits — capping overtime while allowing 30 hours a week of secondary work does not achieve the goal.
On distribution
Use data to identify when and why overtime is occurring. Where staffing shortages drive routine overtime, consider relief officers or adjusted shift overlaps rather than paying indefinitely. Where overtime is necessary, keep it voluntary and evenly distributed rather than concentrated on a subset of officers.
On tooling
Explore scheduling systems that can flag when an officer is about to exceed the limits you have set, track hours, and distribute assignments evenly across a 24/7 operation. Quantify the cost of fatigue in overtime, accidents, and lost productivity when making the case for staffing levels.
Scheduling arithmetic that changes the answer.
A compressed rotation does not reduce headcount
A post staffed around the clock needs 8,760 hours of coverage a year. That total is a property of the post, not of the shift length. Divided by a 2,080-hour work year it comes to 4.21 officers per post before a single hour of leave, training, court, or sick time — and every agency has to add for those.
Shift length changes when officers are present and how many changeovers occur. It does not change the hours that have to be covered. A department expecting a move to 12-hour shifts to let it run leaner is working from arithmetic that does not hold.
Fix the work period before you compare the patterns
Both 12-hour rotations above average 42 hours a week. Measured against a seven-day, 40-hour workweek, that is structural overtime on every officer, every week — which makes the pattern look untenable on cost before it has been evaluated on anything else.
Section 7(k) of the FLSA lets employees engaged in law enforcement be paid overtime on a work period basis instead, where a work period may run from 7 to 28 consecutive days. For a full 28-day period, overtime is owed above 171 hours for law enforcement. For periods of at least 7 but fewer than 28 days the threshold is proportional — a 14-day work period comes to 86 hours.
The exemption is not automatic. The agency has to declare the work period, and the regulations require a payroll notation recording the work period for each employee. Different groups of employees may be given different work periods.
Under a declared 28-day period, the same 42-hour average is a very different number. Comparing rotations against a flat 40-hour week when your agency has declared a 207(k) period will produce a cost comparison that does not reflect what you actually pay.
This is a general description of the exemption, not legal advice. See the Department of Labor’s Fact Sheet #8 and 29 C.F.R. Part 553 Subpart C, linked in the sources below, and consult your own counsel on how it applies to your agency.
Building the candidates, and holding the limits you set.
Snap Schedule 365 will not tell you which rotation to choose. It has no model of officer fatigue and no forecast of what a pattern will cost you. What it does is let you build each candidate and look at it before the decision is made.
- Each rotation is a grid you fill onceA shift pattern is columns for days and rows for crew positions, so a Pitman is four rows by fourteen columns. A schedule plan then says which crew starts on which row and rotates them forward.
- The plan shows how hours fall across crewsWork-hours figures on the plan grid assume one employee per team. They are useful for comparing how evenly a rotation distributes hours — the fairness question a bid cycle will raise — not for total labor cost.
- Overtime is calculated against your declared work periodDaily, work-period, and seventh-consecutive-day thresholds, each with tiers, so a 42-hour average is evaluated the way your agency actually accounts for it.
- The limits you decide on become flagsWork-hour limits, rest requirements, and consecutive-day maximums are configured once and surfaced by schedule validation. They are advisory flags for a supervisor to review, not hard blocks.
Two things to know before you try this
Generation writes to the live schedule and there is no undo. It applies the pattern and nothing else — it does not consult approved time off, availability, or existing assignments, and its delete option is on by default. Build candidate patterns in a trial or a separate account, not on a working schedule.
Review any generated period afterward. Anyone with approved leave in the range will have been scheduled to work it. Conflict flags will show the collisions, but only after the fact.
Seeing the handover gap
Coverage Analysis charts required against scheduled hour by hour for one location and date, at intervals down to 15 minutes. Negative variance in short bursts around a shift change is the handover gap — the window where one assignment ends before the next begins. Narrowing the interval shows exactly how wide it is.
It reports task coverage, so it requires task requirements to be set up. It only reports; it never blocks an assignment.
How each rotation is built.
Mechanics for the individual rotations, including several 24/7 patterns beyond the four compared above.
Questions about police shift patterns.
Which shift length is best for a police department?
The research does not support a single answer, and we are not in a position to give you one. What the randomized experiment found is that 10-hour shifts produced the most sleep, the highest reported quality of work life, and the least overtime, while 12-hour shifts produced lower reported alertness with no quality-of-work-life benefit. It also found no significant differences among the three lengths on work performance, health, or work-family conflict. The National Policing Institute’s published recommendation is to consider four 10-hour days as a strong option for patrol and to implement 12-hour shifts cautiously and voluntarily. How that weighs against your staffing, your geography, your call volume, and what your officers will accept in a bid is a judgment only your agency can make.
Which police shift pattern has the least overtime?
In the Amendola experiment, officers on 8-hour shifts worked more than five times as much overtime as those on 10-hour shifts and more than three times as much as those on 12-hour shifts — so the 10-hour group worked the least, roughly 80 percent less than the 8-hour group. The proposed mechanism is that fewer shift changes mean fewer occasions for an officer to stay late finishing a call, and that overlapping shifts absorb that work inside the scheduled day. Note that the study measured shift length, not specific named rotations.
Will moving to 12-hour shifts let us run with fewer officers?
No. A post covered around the clock requires 8,760 hours of coverage a year regardless of how the shifts are divided. Against a 2,080-hour work year that is 4.21 officers per post before any allowance for leave, training, court, or sick time. Shift length changes when officers are on duty and how many changeovers occur; it does not change the number of hours that must be covered.
How does a 42-hour average rotation work under the FLSA?
It depends on the work period your agency has declared. Against a standard seven-day, 40-hour workweek, a 42-hour average is overtime every week. Section 7(k) of the FLSA lets law enforcement employees be paid overtime on a work period basis, where the period may run from 7 to 28 consecutive days. Over a full 28-day period, overtime is owed above 171 hours; for shorter periods the threshold is proportional, so a 14-day work period comes to 86 hours. The agency must declare the work period and record it in payroll. This is a general description rather than legal advice — see the Department of Labor’s Fact Sheet #8 and 29 C.F.R. Part 553 Subpart C, and consult your own counsel.
Is 12-hour shift fatigue a reason not to use the pattern?
The experiment found that officers on 12-hour shifts reported greater sleepiness and lower alertness than those on 8-hour shifts, and the researchers advised caution on that basis, noting that people underestimate their own fatigue and that accident risk rises with hours worked. But the same study found no significant difference in measured work performance, driving, or reaction time. What was demonstrated is a difference in reported alertness, not a demonstrated decline in performance — a distinction worth preserving when the finding is quoted in a labor-management discussion.
Can Snap Schedule tell us which pattern to adopt?
No. It has no fatigue model and does not forecast what a rotation will cost. It lets you build each candidate rotation as a shift pattern and schedule plan, shows how hours distribute across crews, calculates overtime against your declared work period, and flags assignments that cross work-hour, rest, or consecutive-day limits you have configured. Those flags are advisory for a supervisor to review, not blocks. Schedule generation writes to the live schedule with no undo and does not consult approved time off or availability, so build candidate patterns in a trial or separate account rather than on a working schedule.
Read the originals.
- Amendola, K., Weisburd, D., Hamilton, E., Jones, G., & Slipka, M. The Shift Length Experiment: What We Know About 8-, 10-, and 12-Hour Shifts in Policing. Police Foundation / National Policing Institute, 2012. NIJ award 2005-FS-BX-0057. Also published as “The impact of shift length in policing on performance, health, quality of life, sleep, fatigue, and extra-duty employment,” Journal of Experimental Criminology 7(4), 2011. NIJ listing
- Pearsall, B. “Sleep Disorders, Work Shifts and Officer Wellness.” NIJ Journal 270, June 2012. Summarizes both the shift length experiment and the sleep disorder screening study. Read at NIJ
- National Policing Institute. “Shift Work, Fatigue, and Overtime in Policing: Balancing Officer Wellness and Public Safety.” InFocus, January 2026. Source of the recommended practices reproduced above. Read at NPI
- James, L., et al. “The impact of work shift and fatigue on police officer response in simulated interactions with citizens.” Journal of Experimental Criminology, 2017.
- Shockney, T., et al. “Shiftwork May Lead to Health Problems among Police Officers,” using Buffalo Cardio-Metabolic Occupational Police Stress study data. NIOSH, 2022.
- Vila, B., & Kenney, D. “Tired Cops: The Prevalence and Potential Consequences of Police Fatigue.” NIJ Journal, 2002.
- U.S. Department of Labor, Wage and Hour Division. Fact Sheet #8: Law Enforcement and Fire Protection Employees Under the Fair Labor Standards Act (FLSA). Read at DOL
- 29 C.F.R. Part 553, Subpart C. Fire Protection and Law Enforcement Employees of Public Agencies — including § 553.230 for the maximum-hours table and § 553.224 on declaring work periods.
Once you have chosen, we can help you run it.
Snap Schedule 365 builds rotations, calculates overtime against your declared work period, flags the limits you set, and puts the schedule in officers’ hands. Published pricing and a 30-day trial that starts without a sales call.