Living in the wrong time zone: Elevated risk of traffic fatalities in eccentric time localities

Abstract

Chronobiology research has uncovered a host of maladies linked to social jetlag (SJL), the sleep-disrupting disconnect between solar time and social time. This interdisciplinary study applies chronobiology theory to the potential effect of misaligned time zones on motor-vehicle deaths. In the U.S. 53 million residents live in counties located outside their official time zones’ standard 15° span of longitude, based on degrees west of the prime meridian. We refer to these counties as eccentric time localities (ETLs), all of which lie west of their time zones’ standard western border in the U.S. In contrast, counties within 7.5° of their time zone’s standard geographic center are what we call solar zones. Solar zones do not vary more than 30 minutes from true solar time. ETL residents are forced to rise before dawn, possibly restricting their sleep-time and suppressing both morning and evening zeitgebers that would support their circadian entrainment. Hypothesizing that living in ETLs amplifies social jetlag, data on 417,399 traffic fatalities in the U.S. between 2006 and 2017 from the Fatality Analysis Reporting System (FARS) census were analyzed via GIS mapping and population-data statistics. Road fatalities among residents of solar zones were compared to those living in ETLs within the same official time zone. ETL residents across the U.S. indicated 21.8% higher fatality-rates than solar residents, with a mean of 1286 additional (i.e., unexpected) deaths-per-year. Results support circadian entrainment theory and are consistent with the SJL construct. The socio-political ramifications of these findings are discussed, as well as the subject of best practices when analyzing whole-population data. The authors conclude that the unquestioned rhetoric of time-zone boundaries should be reconsidered in social policy.

Keywords

Social jetlag eccentric time localities chronobiology circadian entrainment dysfunctional social time fatality analysis reporting system census

In 2019, economists Osea Giuntella and Fabrizio Mazzonna made international headlines with research on the health impacts of social jetlag at time-zone boundaries (e.g., Ingraham 2019). Their study was grounded in a body of research led by circadian biologist Till Roenneberg, who since 2003 has done more than anyone else to alert humanity to the unintended consequences of alarm clocks and other modern sleep thieves. Giuntella and Mazzonna (2019) eschewed policy inferences but set the stage for further research on how time-zones may influence quality of life, and how societies could account for this knowledge in social policy.

The present study expands on Giuntella and Mazzonna’s (2019) research on time-zone boundaries to observe how erratic time zones may influence motor-vehicle safety on large scales. Unlike their study, we also consider the rhetorical and policy ramifications of such findings. Grounded in circadian-entrainment theory (e.g., Kantermann 2013), this exploration provides a highly interdisciplinary approach, informed by the fields of communication, geography, and biology. After a review of the salient research literature, we establish the phenomenon of eccentric time localities (ETLs) and how they may be associated with vehicle-fatality rates. The results are described and their implications considered.

Rising awareness of social jetlag

In the first decade of the present century, researchers in chronobiology identified the phenomenon of social jetlag (SJL), defined by Roenneberg and Merrow (2016) as “the difference in sleep timing between free days and workdays (expressed in hours) and is the consequence of the alarm clock” (p. 434). Since the explosion in the use of artificial lighting a century ago, society time has become disconnected from the sun’s movement.

Most creatures on earth are attuned to the 24-h solar day, although individuals come with varying chronotypes, or body clocks. Among humans, chronotypes range from extreme larks (early to rise, early to bed) to extreme owls (late sleepers), with doves in between. Larks wake easily in the morning and fade earlier in the evening compared to owls due to differences in their individual circadian rhythms. Despite individual differences, regulators known as zeitgebers help owls keep time with the solar day as well. Chronobiologists call this phenomenon the process of entrainment (e.g.,Roenneberg and Merrow 2016). For 200,000 years, homo sapiens lived with few distractions from these synchronizing zeitgebers. Thus, owls were not so much at a disadvantage in the morning compared to larks. However, in the last 150 years, industrialization, alarm clocks, the incandescent lamp, shift work, daylight saving time (DST), and other social factors have rivaled the solar clock for influence amongst many cultures and populations. Wright et al. (2013) refer to this modern phenomenon as “the constructed environment” (p. 1554).

Under natural conditions, the exogenous solar day interacts with the endogenous circadian cycles of individuals to form the process of entrainment. This process is delicate, however, and can be disrupted by social influences. The constructed environment now forces most of us into later chronotypes than we would occupy naturally. Later bedtimes become a problem when morning start-times are fixed, as in work and school schedules. The disconnect between solar time and social time leads physiologically to a circadian misalignment, measured as social jetlag, or SJL (Roenneberg et al., 2019).

Roenneberg and Merrow (2016) cite self-rated scales suggesting that 87% of people in wealthy, minority-world countries suffer from social jetlag, which entails eating, working, and exposure to artificial light during hours in which we would naturally be asleep. Consequently, “small insults from chronic circadian misalignment lead to diverse pathologies” (Roenneberg and Merrow 2016: 436). Surveying the evidence for SJL, we find these coinciding disruptors to natural sleep sufficiently pervasive and pernicious to amount to dysfunctional social time, a rhetorical/conceptual contribution to Roenneberg’s SJL construct. If social jetlag is the effect, then dysfunctional social time is a cause.

Nighttime sleep-deprivation correlates with higher tobacco and alcohol use (Raloff, 2006), short-term cognitive decline, and long-term dementia (Roenneberg and Merrow, 2016). Obesity and diabetes are positively associated with SJL. Giuntella and Mazzonna (2019) explain how socially jet-lagged individuals eat later in the evening, when their metabolisms are less active, with a tendency to dine-out where high-fat foods dominate the menu. Roenneberg (2004) notes that we are forced to eat and work when we are tired and sleep when we are hungry.

One possible way to reverse the effects of social jet lag may be the resumption of pre-industrial sleep habits. In a study of modern city dwellers who took a week off to live in the woods, Wright et al. (2013) found that a return to natural-light zeitgebers, without the interference of alarm clocks, seemed to reverse SJL via entrainment: re-synchronizing the human circadian clock back to solar time. The phenomenon of larks-versus-owls diminished, giving all chronotypes improved sleep. Although the study was limited in scope and provided no long-term results, its conclusions are consistent with the SJL construct. The potential causal chain of social jetlag is illustrated in Figure 1.

Figure 1.

Theoretical Causal Chain of Modern Social Jetlag.

Social jetlag has been established as a serious problem not only for shift workers (one-fifth of the U.S. working population), but for the vast majority of all people in the minority world (Roenneberg and Merrow, 2016). What if SJL is then compounded for millions of people because their counties are placed in the wrong time zone?

SJL, time zones, and death

Social jetlag is caused by an array of contributors to dysfunctional social time, including alarm clocks, artificial lighting, Daylight Saving Time (DST), etc., as demonstrated above. DST has come under scrutiny by the European Parliament (Economist, 2019), while some U.S. legislators take the opposite view and wish to make DST permanent (e.g., Rubio and Buchanan 2019). However, the present study targets the problem of flawed time-zone boundaries in the U.S. year-round. As noted by Roenneberg et al. (2007) and Benediktsson and Brunn (2015), time zones are politically drawn approximations of solar time, making them a social construct. In other words, time zones are rhetorical.

Earth’s 360 degrees of longitude are divided into 24 h, or time zones, averaging 15° each. These time-zones are not always straight lines drawn north-to-south, but are socially derived approximations of solar time. In some regions, they are poor approximations due to being drawn erratically. One example is North America’s Eastern Time Zone. On December 21, 2020, sunrise in Marquette, Michigan occurred at 8:30 a.m. Eastern Standard Time (EST). Bangor, Maine is classified in the same time zone, but sunrise there occurred at 7:09 a.m. Whereas Bangor lies appropriately within Eastern Time, Marquette is more than 300 km west of the time zone’s solar edge, making it geographically within Central Time (i.e., −6 h west of Greenwich Mean Time (England), abbreviated as GMT). Therefore, the two cities have far different wake/sleep times each day in relation to sunrise and sunset. Applying Roenneberg’s system of chronobiology, Marquette emerges as the loser. In winter, residents there must awaken long before sunrise to start their day, missing out on zeitgebers that Mainers can take for granted. If time zones were drawn to better match solar time, no location should be much more than 30 min off of solar time. But Marquette is more than an hour of daylight west of the Philadelphia center (1 h, 11 min). We have identified no discussion of the merits of these eccentric boundaries in the U.S., either in the research literature or public debate. Current time-zone boundaries in the U.S. appear to be unconsciously accepted by scholars and policy-makers alike.

The entire state of Michigan falls geographically within what would be the Central Time Zone if it conformed to the standard 15° spans; yet 90% of the state is officially placed in Eastern Time. Their potential circadian misalignment is compounded by another full-hour during DST, which occupies nearly two-thirds of the year for all Americans except residents of Hawaii and most Arizonans. In counties within 7.5° of their time-zone’s solar center, such as the entire Pacific Time zone in the U.S., SJL is theoretically less pronounced than where boundaries are more variable. Erratic time zones affect Eastern, Central, and Mountain Time, as any time-zone map reveals (see Figure 2).

Figure 2.

Solar Zones Versus Eccentric Time Localities (ETLs), With ETL Counties Shown West of Each Black Line. Note. Black lines represent geographic time zone edges matching 15° delineations starting from Greenwich, England (0°). Non-conterminous color west of these straight lines indicates eccentric time localities.

The black lines in Figure 2 represent what would be the edges of each 15° time zone if they were drawn to match the movement of the Sun. Official Mountain-Time counties are colored yellow, but they only roughly match the solar zone occupying the space between the two black lines on the left half of the map. Many are yellow even though they lie west of the far-left black line. Even more counties in the official Central (orange) and Eastern (white) zones are found west of the straight lines that would mark their western edges if the official borders matched their standard 15° spans west of GMT (0° longitude).

No counties in Pacific Time (green) stray more than 7.5° from the solar center, which happens to slice very near Reno, Nevada (120° West of GMT). In Reno, the sun is at its highest point at twelve-noon during the 4 months of standard time each year. In the other three time zones, however, the solar center lies substantially east of the politically drawn mean longitude (see Figure 2). This means that millions of Americans, who live in the western parts of these three time zones, are forced to “lean east.” The result is a misalignment of waking time to morning light, which is consistent with SJL. We describe counties located in the “wrong time zone” as eccentric time localities (ETLs), whereas counties located within 7.5° of their solar center are solar zones. In solar zones, the sun is at its highest point within 30 min of 12:00 noon during standard time.

All ETLs in the United States privilege the east, never the west. Giuntella and Mazzona (2019) confirm that nearly all county-requested exemptions to the Standard Time Act of 1918 moved time-zone boundaries west. As the central meridian of Mountain Time runs through Denver, Colorado (105°), one would expect it to be equidistant from the borders of both Pacific and Central Time. Yet the Mile-High City is only 250 km west of the Central Zone, compared to 770 km east of the Pacific zone, which illustrates the east-facing skew of Mountain Time. Likewise, the Central Time Zone’s meridian dissects New Orleans, Louisiana (90°), which is more than three-times closer to counties observing Eastern Time (420 km) than to the Mountain Time Zone (1380 km). Philadelphia, Pennsylvania (75°) sits on the central meridian for Eastern Time. It is 720 km west of Atlantic Time but 1140 km east of its own time zone’s western edge. Consequently, these three irregular time zones (Pacific Time has no ETLs in the U.S.) contain many ETLs, with millions of people living in them. If misaligned time zones disrupt circadian rhythms in humans, large-scale epidemiological studies could illuminate any resulting negative health and social effects. Such effects would be consistent with theory and research in chronobiology.

In a study of Russia and China by Borisenkov (2011) one’s position within a time zone was found to be a significant predictor of cancer incidence and mortality, with those living in the far western part at greater risk than those living to the east in the same official zone. A study of 11 U.S. states (Gu et al., 2017) likewise indicated significantly higher cancer rates in the western part of the time zone. They suspect these findings may be linked to circadian disruption associated with time-zone misalignment. Consistent with the foundational research on social jetlag, both studies suggest that location within time zones may affect sleep and therefore health, prompting the need for further investigation into the potentially harmful effects of irregular time zones.

Key studies

Two lines of research offer particular salience for the present study: those associating Daylight Saving Time (DST) with automotive fatalities, and Giuntella and Mazzonna’s (2019) study on how one’s placement within U.S. time zones correlates with a variety of health outcomes. DST has served as a point of controversy since its first official use in the early 20^th century. At the same time, auto accidents are a leading cause of death in the Western world and the number-one cause of accidental death in the U.S. at about 100 per day. From a rhetorical perspective, we find it strange how Americans passively accept the carnage on the nation’s roads. If 100 people died each day in airline crashes, it would likely constitute a national emergency.

Previous literature demonstrates mixed findings on the impact of DST on traffic accidents. Coate and Markowitz (2004) concluded that DST provides marginal benefits to traffic safety due to increased ambient light late in the day. Carey and Sarma’s (2017) observed inconclusive results in a review of 24 short-term and long-term DST studies, 17 of which were in the United States. Varughese and Allen (2001) examined 20 years of automobile crash data in the U.S. to demonstrate an increase in automobile crashes the Monday following DST. Fritz et al. (2020) focused on a chronobiological context by quantifying DST, time of day, and time zone effects when exploring all motor vehicle accidents between 2006 and 2017. They found that Spring DST increased automobile fatalities by 6% and a more pronounced fatality rate during morning times further west within a time zone.

Consistent with Fritz et al. (2020), Smith (2016) used large data sets to observe a significant increase in fatal auto accidents in the U.S. during the switch to Daylight Saving Time (DST) each spring. Over a 10-year period, fatal crashes rose 5.6% during the first week of DST and 3.2% over its first month. Smith found no compensating improvement in traffic safety occurring during the “fall-back” when DST ends each autumn. Relevant to SJL, Smith identifies the drowsiness caused by losing an hour of sleep as the primary causative agent, rather than random factors or changes in evening and morning light. Sleep is essential for drivers' alertness and control of a vehicle; therefore, “sleep deprivation likely reduces driving safety,” and “All four tests suggest that the sleep deprivation is driving the increase in fatal crashes” (Smith 2016: 70, 89).

Smith (2016) estimates 30 additional vehicular deaths nationwide in the United States each year attributable to the transition to DST, at an average annual economic cost of $275 million. That amounts to $9.2 million in economic losses-per-death. However, Blincoe et al. (2015), in a study for the U.S. National Highway Safety Administration (NHTSA), estimates that each highway death costs society a smaller amount: $1.4 million. This would mean $42 million in losses each year, still a sizable economic toll.

The reason the DST studies above are salient to our research is not merely that the time-change creates an hour of sleep deprivation for 1 day each year. Rather, the theory-building value comes in the link between sleep disruption and automotive safety. Whereas the adverse effects of DST on driver acuity may last an average of 1 week (Smith 2016), residents of eccentric time localities (ETLs) live under a permanent circadian deficit.

If nearly 90% of all people in minority-world nations live with SJL, as claimed by Roenneberg and Merrow (2016), ETL residents may suffer from a compounded social jetlag merely due to their geographic location. SJL is theoretically exacerbated by the persistent lack of ambient light in the morning when people drive to work or school. This is not just a visibility issue but a problem of entrainment. Although extended evening light will compensate for lost morning light, Smith (2016) notes that daylight affords greater safety for drivers during morning hours than in the evening. Thus, "[t]here is a significant short-term increase in fatal crashes following the spring transition, consistent with a detrimental impact of sleep loss” (Smith 2016: 84). The DST findings above are consistent with Roenneberg’s SJL construct, in which any social practice that forces people into different chronotypes produces detrimental outcomes among multiple physiological systems. If so, the impacts may be compelling, as our GIS analysis reveals that 53 million people in the U.S. live in the “wrong” time zone.

Consistent with research on DST and auto accidents, Giuntella and Mazzonna (2019) observe that people living near each other but on either side of a Pacific-Mountain-Central-Eastern time border experience nearly an hour’s difference in clock-sunrise and sunset times. Reviewing health-data along the 250 miles (400 km) straddling these border regions, residents on the late-sunset side were found to manifest circadian deficits compared to those living across the time border to the west, with adverse health outcomes resulting.

The unlucky residents of the late-sunset side slept an average of 19 fewer minutes per night than those on the early-sunset side, even though they received the same amount of daylight. This deficit equates to 14 missed nights of sleep annually. In late-sunset areas, “the delay of natural light onset would induce a misalignment between biological and social daily rhythms” (p. 213). East-side wages are three percent lower, and obesity rates 5.6% higher, than the west-siders’. The authors also found higher rates of diabetes, cardiovascular diseases, and cancer (breast, prostate, colorectal) on the late-sunset side, all consistent with Roenneberg’s SJL construct and the previously cited cancer studies by Borisenkov (2011) and Gu, et al., (2017).

The data on SJL prompt the question does the presence of appropriate morning light and evening darkness constitute a wonder drug or economic stimulus? As noted by Kantermann (2013), “First, humans—as with most animals—rely on regular light exposure to be entrained and to get sufficient sleep” (emphasis ours; p. 1553). Unfortunately for many in the U.S., daylight is anything but regular. Research supports a relationship between the inhibition of natural zeitgebers and reduced sleep-quality and health. Does it follow that social jetlag affects other factors related to the quality of life? Kantermann (2013) encapsulates why society might benefit from addressing the causes of dysfunctional social time:

Circadian entrainment theory suggests that brighter light in the morning and dimmer light in the evening will shift sleep to earlier hours for most people, a finding that has also been shown experimentally… Furthermore, solely reducing light exposure in the evening hours was also shown to be beneficial to sleep quality (p. 1553).

If Kantermann is correct, eccentric time localities impair sleep quality by their very existence, promoting adverse health and social outcomes. Even more relevant to driving acuity, Roenneberg, et al., (2007) conclude that circadian timing is “enormously influential” on human factors that include alertness and performance (p. 436).

If time-zone misalignment contributes to circadian misalignment, the harmful effects may be substantial. According to Tefft (2018), driver drowsiness is directly involved in seven percent of motor-vehicle crashes but 16% of fatalities, revealing how lethal sleep disruption is on the roads. The U.S. National Transportation Safety Board estimates that fatigue-inducing lapses of attention are a suspected contributor in 40% of all major accidents (Marcus and Rosekind 2017). Tefft (2018) cites a panel of experts who declare that sleep-deprived drivers are unfit to operate a motor vehicle. Even if the elevated risk from misaligned time zones to the average individual is modest, its impact spread over 53 million people living outside their solar time zones could be substantial.

Our research is motivated by the following hypothesis: Counties in eccentric time localities (ETLs) will indicate a meaningfully higher rate of motor-vehicle fatalities than those to their east within the same official time zone. The east is denoted because all ETL counties in the U.S. lie west of the geographic time zone (see Figure 2). This hypothesis could be flawed because ETLs do not experience less light each day than their counterparts living in solar time zones. Moreover, more auto accidents happen in the evening than in the morning (Smith, 2016), so ETLs may provide a net advantage in traffic safety. However, Smith (2016) notes that daylight is more critical to traffic safety in the morning than the evening. All things considered, if Roenneberg’s et al. (2007) SJL construct is sound, we should find higher auto fatality rates in eccentric time localities due to reduced driving acuity, which in turn is affected by abnormal circadian entrainment (see Figure 3).

Figure 3.

Potential Causal Chain From Eccentric Time Localities (ETLs) to Traffic Fatalities.

Methods

As discussed, previous literature has identified correlations between chronobiology and health, building to the SJL construct (e.g., Roenneberg, et al., 2007; Roenneberg and Merrow 2016). The association between DST and automobile fatalities is likewise established (e.g., Smith 2016). The present study compares vehicle-fatality rates in the U.S. by county from 2006 to 2017. The comparison groups are residents of eccentric time localities (ETLs) and those living in solar zones.

GIS analysis

The Geographic Information System (ArcMap 10.4, ESRI, Redlands, California, US) and the National Atlas of the United States (2005) were utilized to retrieve time zones of the 48-contiguous United States, and we used a county map retrieved from Topologically Integrated Geographic Encoding and Referencing system (TIGER). We used a base map of the continental U.S. in ArcMap to extrapolate 15° longitudinal lines that delineate the solar time zones (see Figure 2). In ArcMap, we used Union Tool to assign counties to their official time zones. The select tool identified counties more than 7.5° west of each time zone’s solar meridian, representing ETL counties (see Figure 4). As solar versus ETL delineation lines run through some counties, we define ETLs as those with two-thirds or more of its area outside the solar time zone, with the population area not skewed otherwise. Figure 4 designates solar counties in green and ETLs in yellow.

Figure 4.

U.S. Solar Zones Versus Eccentric Time Localities, With Solar Zones in Green and Eccentric Time Localities in Yellow.

Fatality data

Like Smith (2016), we used the Fatality Analysis Reporting System (FARS) nationwide automobile fatality census provided by the National Highway Transportation Safety Administration (FARS, 2021). The starting point of the data, 2006, was chosen because in that year Indiana joined Daylight Savings Time (DST), eliminating the need for adjustments. Because most of Arizona does not participate in DST, the Grand Canyon State was excluded from analysis. Data from Alaska (extreme width of 44° longitude) and Hawaii (no use of DST) were excluded as well. The range of data utilized was the FARS census from 2006 to 2017.

Vehicle fatality rates (VFRs) were calculated by dividing the total enumerated deaths from 2006 to 2017 by population size for each county. We used 2013 population size estimates, which lie within the 12-years of comprehensive data (https://raw.githubusercontent.com/STAT-JET-ASU/Datasets/master/Instructor/uscounties2013.csv). With a topic such as automobile fatalities, urban primacy is an important geographic factor, with metropolitan areas demonstrating higher raw vehicle fatality numbers. Consequently, we further categorized counties based on census designations to discern if variation in urbanization across time zones would affect the results. Counties were divided into the following statistical areas: metropolitan (at least one urbanized area with >50,000 population), micropolitan (at least one urban cluster of 10,000–50,000 population), and rural (population <10,000; see Figure 5). Means were compared and Cohen’s d_s applied to discern if mean differences were consistent in each census-designation area.

Figure 5.

Metropolitan, Micropolitan, and Rural Statistical Areas in the Contiguous 48 States: Tiger/Line Shapefile 2019

Population-data analysis

This study employs a quantitative lens utilizing Geographic Information Systems (GIS) analytical capabilities, coupled with the complete enumeration survey method (CESM; e.g., Yoshino, et al., 2021). To test the hypothesis of this study, vehicle fatality rates (VFR) in solar counties were compared with those of ETL counties, with all results reported without manipulation of data or the removal of “outliers.” The present study benefits from the availability of population data, that is, a complete census of traffic-fatality data over 12 years, rather than relying on samples of that population. This analysis entails a careful review of the descriptive data itself, rather than inferential-statistical estimates. This approach is appealing because population data involve no standard error.

Salient statistics for the analysis of population data include the dependent-variable means across salient independent variables, reporting of the pooled standard deviations, and use of Cohen’s d_s and Pearson’s r to estimate the effect size. Cohen’s d_s facilitates the comparison of groups of different sizes, such as the populations of solar time zones and ETLs. The resulting statistic represents the number of standard deviations (SD) separating the groups, where 0.50 indicates a difference of one-half SD. A Pearson’s r statistic was utilized to measure the correlative relationship between solar-ETL status and the VFR. The predetermined Alpha α level for the Pearson’s r statistic was 0.05.

When all individual cases are enumerated, no standard errors need to be mitigated by tools such as probability values (p) or the removal of outliers to achieve assumptions of a normal distribution (Alexander, 2015). Confidence in the FARS dataset is aided by the extensive quality controls and training provided to analysts at the National Highway Traffic Safety Administration (NHTSA) each year (Anders Longthorne, Mathematical Statistician, NHTSA, personal communication, December 3, 2020).

Cowger (1984) further identifies the problem of “scientific ritualism,” in which the same, customary research tools are used regardless of their goodness-of-fit in order to “lend an air of scientific objectivity to conclusions” (p. 360). Specifically, the author affirms that claims of statistical significance are not always coextensive with practical significance, substantive significance, theoretical significance, or magnitude. In other words, a p-value below 0.05 does not always signify an important result, and vice versa. When researching population data, descriptive statistics advance accuracy, freedom from bias, parsimony, and they eliminate sampling error.

Results

Data from the Fatality Analysis Reporting System (FARS, 2021) between 2006 and 2017 indicated 430,593 traffic fatalities in the United States, an average of 35,882 per year. Subtracting Alaska, Hawaii, and Arizona per the methods above, the enumerated deaths total 417,399 over the 12-year period across 3094 counties. The highest number of deaths, not accounting for population density, was Los Angeles County, California, with an average of 639 deaths-per-year. Loving County, Texas (an ETL) had the worst vehicle fatality rate (VFR). Over 12 years, it lost 11 out of 95 residents to fatal crashes (12%). At the other extreme, Buena Vista County, Virginia and McPherson County, Nebraska had zero vehicle fatalities over the entire 12-year period.

The Pacific time zone indicated the lowest mean VFR among the four contiguous U.S. time zones over 12 years, with a rate of 0.001059. The Central zone had the highest mean VFR at 0.001624, or 53% higher. Eastern Time: 0.001280 and Mountain Time: 0.001538 fell in between (Table 1). The Central zone’s VFR means one out of every 615 residents died in automotive crashes over the 12-year period, or one in 7387 per-year.

Table 1.

Vehicle fatality rates (VFR) by official time zone (2006–2017).

Time zone	#Counties	Vehicle fatality rates	Total deaths	Total population
Eastern	1152	0.001280	191,200	149,332,809
Central	1493	0.001624	148,679	91,529,476
Mountain	290	0.001538	22,139	14,399,283
Pacific	159	0.001059	55,381	52,317,788
Total	3094	0.001357	417,399	307,579,356

This study probes whether ETLs indicate higher vehicle fatality rates (VFRs) than the solar zones to their east, using measures of practical significance. The mean VFR among solar zones in these 47 states was 0.0013074, compared to the ETL rate of 0.001592 (Table 2). In a hypothetical city of 100,000 people, ETL residents averaged 159 total deaths between 2006 and 2017, compared to 130 deaths in solar-zones: a difference of 21.8%. Cohen’s d_s provides an effect-size estimate of 0.823 (in which 1.0 = one standard deviation). Pearson’s r statistic (counting each county equally) indicated a correlation between solar-ETL status and VFR of 0.109 (p < 0.01)

Table 2.

Vehicle fatality rates (VFR) comparing all solar time zones versus all eccentric time localities (2006–2017).

Class	Vehicle fatality rates	Total deaths	Total population	Pct. Worse	SD	Cohen’s d_s	Pearson’s R
Solar	0.001307	332,173	254,068,761		0.22250
Eccentric Time Localities	0.001592	85,226	53,510,595	21.8%	0.54183
Total	0.001357	417,399	307,579,346		0.34664	0.823	0.109 (p < .01)

Viewing each time zone independently, the counties with ETLs in all three time zones indicated higher VFRs than their solar counterparts (see Table 3). In the Eastern zone, ETLs presented a death-rate 23.8% higher than their solar-zone. Central-Time ETLs experienced 17.7% more deaths. In Mountain Time, ETL deaths exceeded solar-zone deaths by 26.5%. Pacific Time, with no ETLs, had the lowest VFR of any region. ETLs overall carried a death-rate 50.4% higher than the Pacific zone.

Table 3.

Deaths and unexpected eccentric time localities (ETL) deaths by time zone; each time zone divided into solar zones and ETLs (2006–2017).

Time zone	Class	Vehicle fatality rat	Total deaths	Total population	Unexpected deaths	Pct. Worse
Eastern	Solar	0.001205	132,722	110,133,372
	ETL	0.001492	58,478	39,199,427	11,239	23.8%
Central	Solar	0.001585	124,890	78,783,002
	ETL	0.001866	23,789	12,746,474	3583	17.7%
Mountain	Solar	0.001494	19,180	12,834,599
	ETL	0.001891	2959	1,564,684	621	26.5%
Pacific (all solar)		0.001059	55,381	52,317,788
Total		0.001357	417,399	307,579,346	+15,443 in ETLs	21.8%

ETL = eccentric time localities

By comparing death rates in solar-vs-ETL zones, we can calculate the number of unexpected deaths in ETLs (also in Table 3). If ETLs in Eastern Time had the same death rate as their solar-zone counterparts, they would have experienced 11,239 fewer deaths over 12 years than they did. Central-ETL deaths would have been 3584 fewer, and the Mountain zone would have counted 621 fewer fatalities. Overall, ETLs experienced 15,443 unexpected deaths over 12 years.

To test for urban primacy, vehicle-fatality rates were compared according to census-designation areas across time-zone classes (see Table 4). Danger decreased as population density rose, with rural areas’ VFR the highest at 0.003111, micropolitan areas’ indicating 0.002182, and metropolitan counties the lowest at 0.001148. In the 1156 metropolitan counties, the ETL death rate of 0.001306 was 16.8% percent higher than the solar rate of 0.001118 (Cohen’s d_s = 0.101). The 650 micropolitan counties did not adhere to the same pattern, as the solar death rate of 0.002193 was 1.9% higher than the ETL figure of 0.002151 (Cohen’s d_s = 0.016). In 1288 rural counties, ETLs indicated a VFR of 0.003316, which was 9.0% higher than the solar mean of 0.003042 (Cohen’s d_s = 0.065).

Table 4.

Vehicle fatality rates (VFR) by solar zones versus eccentric time localities across census designations (2006–2017).

Designation	#Counties	Solar VFR	ETL VFR	SD	ETL Pct. Worse	Cohen’s d_s
Metropolitan	1156	0.001118	0.001306	0.18664	16.8%	0.100
Micropolitan	650	0.002193	0.002151	0.23667	–1.9%	0.017
Rural	1288	0.003042	0.003316	0.42034	9.0%	0.060
Total	3094	0.001307	0.001593	0.45440	21.8%

ETL = eccentric time localities; VFR = vehicle fatality rates.

Discussion

Amarillo, TX, and Nashville, TN, share the U.S. Central time zone. Children in Amarillo, an eccentric time locality (ETL), begin their school day in the dark for most of the school year. To wit: an alarm clock set for “7:00” am in Amarillo means a solar wake-time of 5:13 a.m. for two-thirds of the year (DST), and 6:13 a.m. during the darkest months (i.e., standard time). Nashville’s position brings advantageous zeitgebers for optimized sleep, more daylight for morning drive-times, and a bright start of the day for most of the year. While this difference may appear inconsequential to laypeople, the rising field of chronobiology indicates otherwise. Our findings on road fatalities in the United States support circadian-entrainment theory. ETL counties experienced higher automobile fatality rates in each applicable time zone among residents of major cities and rural areas alike. Living in eccentric-time localities may contribute to social jetlag.

The mean vehicle-fatality rate (VFR) of 0.001592 in ETL counties was 21.8% higher than the solar VFR of 0.001307, leading to a rejection of the H₀. If ETLs had the same death rate as solar areas, they would have experienced fewer than 70,000 deaths over 12 years. Instead the dead numbered above 85,000, or 15,443 unexpected deaths in ETL regions. As found in Table 3, Eastern Time’s ETLs hold a mean of 936 unexpected deaths annually, along with 298-per-year in Central Time, and 51 in the sparsely populated Mountain zone. These three time zones house four-fifths of the U.S. population.

Nationally, ETLs experience a mean of 1285 unexpected deaths-per-year. This number far exceeds Smith’s (2016) finding of 30 additional deaths-per-year attributable to Daylight Saving Time across the United States. Using the conservative economic-cost formula of $1.4 million per-highway-death used by Blincoe, et al., (2015), this means $1.8 billion in annual economic losses from excess U.S. road fatalities are associated with living in the wrong time zone. This monetary figure helps illuminate the practical significance of ETLs and vehicle fatalities.

We attribute the higher fatality rate in ETLs to circadian disruption along with imposed darkness during the morning-commute hours. Eastern Time’s ETLs lead the results with 936 unexpected deaths-per-year (+23.8%). These residents should rightly be placed officially in (solar) Central Time, that is, −6 h west of GMT. Mountain Time indicates an even greater risk among ETL residents (+26.5%), with Central ETLs (+17.7%) less extreme but likewise unmistakable. Pacific Time, with a VFR 50.4% below that of the combined ETLs, may benefit from the complete absence of eccentric time localities. We hold that an objective observer would characterize the totality of these figures as shocking. Regarding practical significance, if any drug carried a 21.8% higher risk of sudden death than its competitors, it would be taken off the market immediately by the U.S. Food and Drug Administration. The Cohen’s d_s estimated effect-size of 0.823 is strong, especially for a dependent variable as complex as traffic deaths. It means the ETL death-rate is nearly one standard deviation above the solar rate. This figure gains further salience by the 53 million people who live in these eccentric time localities.

A final measure of practical significance is the number of years of life lost to vehicle fatalities (Y LL). According to Webb (2018), vehicle accidents are a leading cause of death in the United States, ranking 13th. In future years of life lost (Y LL), however, the roads are even worse: ranking seventh. The difference is because, compared to other leading causes of death, traffic fatalities disproportionately affect young people—who have more years to lose. Webb reports that the average traffic fatality costs 38 years, 5 months of life. If we can assume ETL deaths manifest the same mean age as other road fatalities (i.e., average death at age 40), this means 49,507 future years of life are terminated each year in ETL crashes above what one would expect if ETLs were not associated with elevated risk.

We found that automobile fatalities occur at substantially higher rates in ETLs than solar zones in the U.S. We did not consider other factors known to impact automobile fatalities, such as variation in speed limits, drunk-driving, commute times, and road quality. While we do not minimize such factors, we believe the census approach provided by the FARS data (2006–2017) sufficiently enumerates U.S. automobile fatalities to associate vehicle-fatalities and time-zone localities.

Research and policy recommendations

This study provides support for circadian entrainment theory and the SJL construct (e.g., Borisenkov, 2011; Gu, et al., 2017; Roenneberg and Merrow, 2016). These findings also prompt several research and policy implications for consideration. First, since daylight hours during school start times are a function of time-zone location, we recommend research on the effects of ETLs on educational outcomes. Second, our methods would be of benefit in further research on other SJL-related maladies in ETLs, such as disparities in wages, obesity, cancer, and other health effects (see Giuntella and Mazzona, 2019). Third, studies of traffic fatalities and SJL should be applied to eccentric time zones worldwide, which are likewise prevalent in Europe, for example (Roenneberg, et al., 2019). The growing body of knowledge of dysfunctional social time should continue or even accelerate. Finally, future research should account for covariates on vehicle fatalities besides urban primacy, such as age, race, gender, and education. Although these data are not available in the nationwide FARS census, one could attempt to draw a representative sample and inferentially test a number of demographic variables.

These results contain important policy ramifications as well. As circadian-entrainment theory addresses a multidimensional problem impacting health, safety, economics, social behavior, and education, the United States should embrace more solar time and less eccentric time by the simple policy measure of moving most time-zone boundaries eastward. As Wright et al. (2013) suggests, circadian disruption and its harmful effects may be reversible. If counties choose not to be intentional about their time zones, school districts and businesses within ETLs should consider later school and work start-times to mitigate their detrimental effects. For example, they could open at 8:30 a.m. rather than 8:00. Policymakers should stop passively accepting the misaligned time zones they inherited from a century past, and recognize them for the socially engineered, rhetorical constructs that they are.

The rhetorical choices in naming time zones are another fruitful area of reform. The vast Central time zone should not encompass the entire center of the continent (which it nearly does) because it is one of four contiguous-U.S. time zones, not three. Central Time would be better termed the “East Central” zone. Likewise, the “Mountain” zone is misnamed. The actual Rocky Mountains are never more than 500 km wide in the contiguous United States, yet this time zone should span a full 1665 km at its median U.S. latitude of 40° north. Consequently, more than half of Texas' land area should be in the “Mountain” zone. Understandably, Texans do not identify as “mountain” people; therefore, this time zone should be called “West-Central” time and expand hundreds of kilometers east. That is, if policymakers care about reducing cancer rates (e.g., Gu, et al., 2017), increasing worker-productivity (Roenneberg and Merrow, 2016), and eliminating hundreds of preventable motor-vehicle fatalities each year. Maladaptive time zones can now be counted among alarm clocks, electric lights, DST, and shift work as a contributor to dysfunctional social time.

Our policy inferences are reinforced by Roenneberg and Merrow’s (2016) statement, “The path forward, then, would involve … protocols to detect sub-optimal entrainment … and the use of zeitgebers to return to optimal entrainment” (emphasis ours, p. 439). Although these authors were speaking on the individual-patient level, our study recommends a return to zeitgebers on a macro-level for millions, in the form of appropriate time-zone boundaries that optimize both bed-and-wake times. Changes in state and/or federal commerce policy can shift millions more Americans into their appropriate time zones at no cost to individuals. At the very least, ETL residents should be made aware of their elevated risk of death on the roads.

These recommendations reinforce Roenneberg et al.’s (2019) more recent research calling for reform of Europe’s time-zones, which are even more eccentric than the United States’. The authors recommend redrawing time zones to reflect solar time better, as well as terminating DST: “Under such adjustment, social (local) clock time will match sun clock time and therefore body clock time most closely” (emphases in original; p. 9). Those in the U.S. who would make DST permanent (e.g., Rubio and Buchanan, 2019) clash with thousands of years of human chronobiology and a rising tide of 21^st-century research. It is not only the number of hours of sleep that matter, but when those hours take place relative to the sun’s position.

A final contribution of this study is to address the problem of ritualism in scientific and social-science research as noted above by Cowger (1984), by providing a parsimonious and appropriate research design for the analysis of whole-population data. Alexander (2015) asks, “Since we have complete data on the population why do summary statistics not suffice, without confidence or credible intervals”? (p. 2). Just as policymakers unconsciously accept maladaptive time zones, some scholars passively apply inferential statistics without regard to their appropriateness (Alexander, 2015). According to Wood (2016 n. pag.), “Obviously if you’ve got the time and energy and resources to study the whole population you should do so because then you get the whole answer and don’t need to mess around with p values or confidence intervals etc. This must be more rigorous. There is obviously no good reason to ignore data you can get!” Better to use parsimonious analytical tools that permit the fully enumerated data to reveal themselves. Perhaps this study can provide heuristic value for other scholars to better analyze the plethora of comprehensive health and social data generated by national governments each year.

Limitations

This study is limited by the complex series of factors that lead to road accidents and fatalities. We do not claim misaligned time zones are the most important factor in road fatalities, due to the influence of speeding, drunk driving, distracted driving, and other causal complexities. Although we accounted for urban primacy, the FARS Census does not include factors such as ethnicity, gender, trip-length, driving speed, or education-level. However, circadian impairment on the individual level has been documented to be an important factor in road fatalities in a number of studies (e.g., Tefft 2018). Our research implicates the role of misaligned time zones in the U.S. as a macro-contributor to these tragedies, and as a potential constituent of social jetlag itself. However, we did not assess social jetlag per se in this study. Therefore, the potential association between ETLs and social jetlag is circumstantial.

A second limitation of the current study is that Solar-vs.-ETL differences were not consistent across census designations in the way they were across time zones. Although metropolitan and rural counties both indicated a sizable disadvantage for ETL residents, micropolitan areas had similar death rates for the two groups. Therefore, residing in solar localities did not provide traffic-safety protection for these residents. As a middle designation, the micropolitan area was added by the U.S. Census Bureau only in 2003 (Smith, 2014), and it may lack the same clear operational definition as the rural and metropolitan designations. For example, communities closer to the population designation’s lower limits of 10,000–25,000 resemble rural patterns, whereas 25,000–50,000 likely best reflects the intention of Micropolitan Statistical Area Census designations. We concede this limitation points to a need for further investigation, but the overall differences between ETLs and solar zones remain substantial.

Conclusion

Roenneberg and Merrow (2005) recognized the heuristic nature of circadian research as “a remarkable example of interdisciplinarity, unraveling the complex mechanisms that underlie a ubiquitous biological programme” (p. 965). Our interdisciplinary study has detected maladaptive time zones and proposes a simple but robust policy solution: move misaligned time-zone boundaries east. Wright, et al., (2013) summarize the widespread and multi-layered problem that we have termed dysfunctional social time:

[I]f human circadian and sleep timing was in synchrony with the natural light-dark cycle, the circadian low point in brain arousal would move to before the end of the sleep episode, making it easier to awaken in the morning. Related, the earlier timing of the melatonin onset after exposure to natural light would promote earlier bedtimes, whereas the later evening timing of melatonin onset and light exposure in the electrical light-constructed environment would promote brain arousal and could contribute to later bedtimes and disturbed sleep (p. 1555).

Wright’s analysis indirectly supports the value of appropriately drawn time zone boundaries. People in solar zones, such as Denver and New Orleans, may wake more easily each morning and have higher-quality sleep. Americans in eccentric time localities, all living west of their solar zones, face a serious concern. The potential harm is magnified during Daylight Saving Time, which for nearly eight months-per-year places them an additional hour west of their solar time zones.

This study has identified an elevated risk of traffic fatalities in eccentric time localities, supporting the growing body of knowledge in circadian-entrainment theory. We hypothesized the existence of a phenomenon based on established theory. We searched for it, and found it. Therefore, we agree with Roenneberg (2004), who argues, “It is time that the profound knowledge of clock research reaches everyone in our society” (p. 785). Interdisciplinary research such as this study can support theory-building in communication, biology, and geography while sparking better-informed social policy.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Jeffery Gentry

References

Alexander

(2015) What’s more general than a whole population? Emerging Themes in Epidemiology 12(1): 1–5. DOI: 10.1186/s12982-015-0029-4

Benediktsson

Brunn

(2015) Time Zone Politics and Challenges of Globalisation. Tijdschrift voor economische en sociale geografie 106(3): 276–290.

Blincoe

Miller

Zaloshnja

, et al. (2015) The Economic and Societal Impact of Motor Vehicle Crashes, 2010. (Revised) (Report No. DOT HS 812 013). Washington, DC: National Highway Traffic Safety Administration.

Borisenkov

(2011) Latitude of residence and position in time zone are predictors of cancer incidence, cancer mortality, and life expectancy at birth. Chronobiology International 28(2): 155–162.

Carey

Sarma

(2017) Impact of daylight saving time on road traffic collision risk: a systematic review. BMJ Open 7(6): 1–14.

Coate

Markowitz

(2004) The effects of daylight and daylight saving time on US pedestrian fatalities and motor vehicle occupant fatalities. Accident Analysis & Prevention 36(3): 351–357.

Cowger

(1984) Statistical significance tests: scientific ritualism or scientific method? Social Service Review 58(3): 358–372.

Economist (2019) Summertime ruse. 431(9137): 43–44.

Fatality Analysis Reporting System (FARS) (2021) National Highway Traffic Safety Administration. Washington, DC: United States Department of Transportation. Retrieved from https://www.nhtsa.gov/research-data/fatality-analysis-reporting-system-fars

10.

Fritz

VoPham

Wright

Jr Vetter

(2020) A chronobiological evaluation of the acute effects of daylight saving time on traffic accident risk. Current Biology 30(4): 729–735.

11.

Giuntella

Mazzonna

(2019) Sunset time and the economic effects of social jetlag: evidence from US time zone borders. Journal of Health Economics 65: 210–226. DOI: 10.1016/j.jhealeco.2019.03.007

12.

Devesa

, et al. (2017) Longitude position in a time zone and cancer risk in the United States. Cancer Epidemiology and Prevention Biomarkers 26(8): 1306–1311.

13.

Ingraham

(2019) How Living on the Wrong Side of a Time Zone Can Be Hazardous to Your Health. Washington, D.C: Washingtonpost.com. Retrieved from https://link.gale.com/apps/doc/A582948618/AONE?u=nm_a_enmu&sid=AONE&xid=54754117

14.

Kantermann

(2013) Circadian biology: sleep-styles shaped by light-styles. Current Biology 23(16): R689–R690. DOI: 10.1016/j.cub.2013.06.065

15.

Marcus

Rosekind

(2017) Fatigue in transportation: NTSB investigations and safety recommendations. Injury Prevention 23(4): 232–238.

16.

Raloff

(2006) Social jetlag: Need a smoke? Science News 170(6): 94. Retrieved from https://search-ebscohost-com-443.web.bisu.edu.cn/login.aspx?direct=true&db=edsjsr&AN=edsjsr.4017060&site=eds-live&scope=site

17.

Roenneberg

(2004) Trying to beat the clock. Nature 427(6977): 784–785. DOI: 10.1038/427784a

, et al. (2007) Epidemiology of the human circadian clock. Sleep Medicine Reviews 11(6): 429–438. DOI: 10.1016/j.smrv.2007.07.005

19.

Roenneberg

Merrow

(2005) Timeline: Circadian clocks—the fall and rise of physiology. Nature Reviews Molecular Cell Biology 6(12): 965–971. DOI: 10.1038/nrm1766

20.

Roenneberg

Merrow

(2016) The circadian clock and human health. Current Biology 26(10): R432–R443. DOI: 10.1016/j.cub.2016.04.011

(2019) Daylight saving time and artificial time zones – a battle between biological and social times. Frontiers in Physiology 10(944): 1–12. DOI: 10.3389/fphys.2019.00944

22.

Rubio

Buchanan

(2019) Make Daylight Saving Time Permanent. Washington, D.C: Washington Post. Retrieved from https://link.gale.com/apps/doc/A578653368/AONE?u=nm_a_enmu&sid=AONE&xid=5e033006

23.

Smith

(2016) Spring forward at your own risk: daylight saving time and fatal vehicle crashes. American Economic Journal: Applied Economics 8(2): 65. Retrieved from https://search-ebscohost-com-443.web.bisu.edu.cn/login.aspx?direct=true&db=edsjsr&AN=edsjsr.24739102&site=eds-live&scope=site

24.

Smith

(2014) Examining the geography of newly incorporated municipalities (NIMs) in micropolitan North Carolina, 1990–2010. Southeastern Geographer 54(4): 384–405.

25.

Tefft

(2018) Acute sleep deprivation and culpable motor vehicle crash involvement. Sleep: Journal of Sleep and Sleep Disorders Research 41(10): 1–11.

26.

Varughese

Allen

(2001) Fatal accidents following changes in daylight savings time: the American experience. Sleep Medicine 2(1): 31–36.

27.

Webb

(2018) Motor Vehicle Traffic Crashes as a Leading Cause of Death in the United States, 2015 (Traffic Safety Facts Crash•Stats. Report No. DOT HS 812 499). Washington, DC: National Highway Traffic Safety Administration.

28.

Wood

(2016) Re: What conditions justify studying the entire population instead of selecting a sample? Retrieved from https://www.researchgate.net/post/What-conditions-justify-studying-the-entire-population-instead-of-selecting-a-sample/573590b05b4952184e44199e/citation/download

29.

Wright

Jr McHill

Birks

, et al. (2013) Entrainment of the human circadian clock to the natural light-dark cycle. Current Biology 23(16): 1554–1558. Retrieved from https://search-ebscohost-com-443.web.bisu.edu.cn/login.aspx?direct=true&db=bxh&AN=BACD201300441122&site=eds-live&scope=site

30.

Yoshino

Kumon

Mori

, et al. (2021) Effects of AGXT2 variants on blood pressure and blood sugar among 750 older Japanese subjects recruited by the complete enumeration survey method. BMC Genomics 22(1): 1–10. DOI: 10.1186/s12864-021-07612-3