A survey run for a national research project on work zone lighting produced an inadvertently useful list. Of 52 state agencies contacted, 33 responded and 28 reported doing some construction or maintenance at night. Ranked by how often they were performed after dark, the most common construction tasks were resurfacing, barrier wall placement, milling and pavement marking. The least preferred were traffic signaling, lighting systems, landscaping, riprap and sidewalks. In the maintenance list, earthwork sat near the bottom.
Nothing in that ranking tracks how dangerous the work is, or how much traffic it disrupts. It tracks one property: whether the job can be started, finished and cleared inside a single window, leaving the road fit to carry the morning peak.
The rule that produces the list
The standing procedure for assessing night work states two conditions that must normally be met for the night shift to offer any advantage at all: reduced traffic volumes, and easy setup and removal of the traffic control pattern on a nightly basis. Then it closes the door on the marginal cases. If the construction operation must occupy the roadway for more than several hours each night, or if the temporary traffic pattern requires too great an effort to deploy and remove, no advantage is gained, and the night work option normally should not be considered further.
That is a stricter posture than the practice suggests. The same report instructs that because night work entails a number of inherent disadvantages, other options capable of addressing the project’s needs should be fully explored before night work is considered.
The candidates it names are the ones that pass the second test: pavement overlays and repairs, bridge work that can progress in nightly increments, and other activities that encroach on travel lanes but can be started and stopped each night. Bridge deck replacement and full-depth pavement reconstruction typically require extended work periods, and the report allows them onto the night shift only in a different configuration, where enough lanes stay open for peak demand and additional lanes close after dark for access.
This is also why an apparently idle work zone is often not evidence of laziness. A traffic pattern too expensive to remove and rebuild nightly is precisely the case in which the guidance says the night option earns nothing, so the cones stay and the shift structure changes instead.
The exposure argument is weaker than the intuition behind it
The usual case for working at night is that fewer vehicles pass the crew, so fewer crashes happen. The first half is arithmetic. The second half is an empirical claim, and the federal work zone economics guidance reports research that complicates it.
Numerous studies, that guidance says, indicate pre-work zone crash rates are likely to increase by 20 to 70 percent when a work zone is in place. For the specific case of active work behind a temporary lane closure on freeways and expressways, it cites a 2008 study by Ullman and colleagues, published as NCHRP Report 627, finding that crash risk for a motorist traveling through the zone increased by about 66 percent during the day and by 61 percent at night. The same passage warns that the actual change varied significantly once the data were separated by time of work and by work condition, and elsewhere notes there are no statistically accepted values for work zone crash modification factors because they vary from study to study.
Take that seriously and the safety case changes shape: for any one driver moving past the work, the danger barely drops at night. What night work reduces is the number of drivers who pass, and the hours they spend queued, which is a claim about aggregate exposure rather than about personal risk.
Against that sits a driving population that is worse on the variables that matter. The assessment procedure is blunt about it. Reduced visibility makes both the driving task and the work task harder. Drivers, pedestrians and workers are generally less alert, the incidence of impaired drivers and pedestrians is often higher, and traffic speeds are higher too. Enhanced traffic controls are generally required at night to compensate, at an increase in project cost and duration. Driver anger and frustration may fall because there are fewer delays, while fatigue and impairment become the greater concern. Countermeasure design for those conditions is the subject of work zone crashes.
National counts cannot settle the question either way. FHWA’s work zone statistics, drawn from NHTSA’s Fatality Analysis Reporting System as reported in April 2024, put work zone fatalities at 963 in 2021 and 891 in 2022, with highway worker occupational fatalities at road construction sites of 108 and 94 in those years. The tabulation carries no split between day and night shifts, so the exposure argument that justifies moving a job into the dark cannot be checked against it.
The variable that governs the shift is the one nobody standardized
Illumination controls safety, productivity and quality together on a night job, and the research team that studied it wrote the sentence that should have ended the argument decades ago: although lighting is the single most important factor in nighttime construction, it is the least studied.
Their account of practice is worse than a gap. Where state specifications addressed lighting at all, requirements were minimally defined: a minimum intensity level of 54 lux, or 5 foot-candles, described as “sufficient light to permit good workmanship and proper inspection.” Most of those specifications, the report says, are not only inadequate but also not standardized, and decisions about work zone lighting are left to the discretion of the site engineer and the contractor. Without criteria for average illumination and uniformity, lighting systems cannot be designed satisfactorily.
Field visits found what that produces. In one case a light plant sat close to the travel lane facing oncoming traffic, producing nearly blinding disability glare for motorists. In others the plants were positioned so as to throw a shadow zone across the work, and tasks were performed in negative contrast rather than positive. Lights on some compaction rollers were not mounted high enough, so most of the light fell on the wheels instead of the pavement. Lighting design and provision, the researchers concluded, were essentially based on the contractor’s discretion, and in some cases little or no thought had been given to the location, positioning and orientation of the light plants.
The guidelines the study produced sort work into three illuminance categories. Category I, 54 lux or 5 foot-candles, covers general illumination in the work zone where crew movement is expected, along with low-accuracy tasks, slow-moving equipment and large objects. Category II, 108 lux or 10 foot-candles, covers illumination on and around construction equipment and the visual tasks associated with it, resurfacing among them. Category III, 216 lux or 20 foot-candles, is for tasks of higher visual difficulty demanding increased attention, such as crack filling, critical connections, maintenance of electrical devices, or moving machinery.
Read the ordering closely. The levels climb with visual difficulty, not with hazard, which puts the general area where crews walk among operating equipment at the bottom of the scale. Glare control was handled as its own set of countermeasures rather than as a byproduct of adding light, for the obvious reason that every luminaire aimed at a crew is aimed somewhere, and the motorist is often where.
Quality runs in two directions after dark
One technical argument for the night shift is real and specific: cooler night temperatures may enhance the quality of concrete placed and finished at night. Anyone who has watched a July afternoon placement fight evaporation and early set knows what is being avoided.
The losses are just as concrete. Reduced visibility and greater difficulty communicating with supervisors or technical support staff are named as night factors that degrade both productivity and the quality of the work, with longer setup and takedown times dragging on output as well. Set against that, reduced interference from traffic and longer uninterrupted shifts can raise productivity outright, so the net is a project fact rather than a general one.
Inspection is where the two problems meet. Inadequate lighting results in problems with proper inspection, and the lighting study is specific about which defects go unseen: shadows, tack spread and asphalt droppings cannot be properly controlled. Those are exactly the judgments an inspector makes by eye, at the illuminance category the guidelines rank hardest, at three in the morning, with the person who could answer a borderline question asleep.
Who pays, and who is asked
The ledger tilts before anyone opens it. On the construction side, the assessment procedure lists overtime and night-premium pay, lighting expense, added traffic control costs and increased material costs, all of which make night work theoretically more expensive to build. On the other side sits work zone road user cost, which FHWA defines as the additional costs borne by motorists and the community at large as a result of work zone activity, monetized through delay, vehicle operating, crash and emission costs. Move a closure to an hour when few people are driving and the largest of those components mostly evaporates. The environmental line points the same way, because emissions and fuel burn fall when congestion, delays and stops fall, which is the mechanism examined in construction equipment emissions.
Who captures that saving depends on the contract. Lane rental fees, schedule-focused contracting strategies and the payment mechanisms compared in design-build against design-bid-build decide whether finishing before dawn is worth anything to the firm doing it, and wider coverage of the field side of this work sits across road construction coverage.
Then there is the column with no unit. The same procedure records that night work may disrupt normal sleep patterns and produce sleep deprivation or poor quality sleep, that normal family and social activity is disrupted, and that the inability to attend community activities or family events may create serious domestic problems. All of it is filed under social factors, weighed qualitatively, alongside a delay figure calculated to the cent. The federal road user cost guidance concedes a version of the same limit on its own side of the fence, treating noise and local community effects as qualitative because no generalized method exists to monetize them.
None of that argues against night work, which on a congested urban corridor is often the least bad option available. It argues for the one instruction in the guidance that is easiest to skip. Before a job moves into the dark, other options capable of meeting the project’s needs are supposed to have been fully explored, and that exploration is a document. Any agency that can produce it has made the trade deliberately. Any agency that cannot has made it by default, on behalf of people who will not be reading the report.