Preferences for Daylight Saving Time meet fundamentals of human physiology: correlations in the 2018 Public Consultation by the European Commission =================================================================================================================================================== * José María Martín-Olalla * Jorge Mira ## Abstract We search for correlations in the preferences for seasonal time arrangements across different countries. With this purpose we analyze the public consultation on summertime arrangements (Daylight Saving Time) launched by the European Commission in 2018 (the largest sample size study of this kind, 4.55 million replies, median shares of population per Member State 0.366%). We also analyze the starting point of human activity in Europe from the Harmonized European Time Use Survey, also the largest scale survey to assess the budget of time in modern societies. We find that the shares of population willing to abolish the seasonal regulations of time increases with increasing latitudes and with delayed winter sunrise times. Notably the distance of the winter sunrise to the starting point of human activity plays a determining role in predicting the preferences (*R*2 = 0.7; *p* = 5 × 10−5; *N* = 17 Member States). We provide a simple physiological rationale for this finding. Our results challenge the current prevailing point of view on clock regulations that associates them to energy saving only, and unveil the physiological roots of these regulations. Keywords * observational studies * quantile regression * statistical modelling * bivariate regression * DST * summer time * society * latitude * sleep deprivation * spring transition * Europe * America * season ## 1. INTRODUCTION Daylight Saving Time (DST) or summertime arrangements is the practice by which many Extratropical (chiefly Western) societies bound to pre-set schedules have successfully regulated their starting times through seasons. The practice consists in a one-hour delay of clock time in autumn followed by a backward one-hour advance of clock time in spring. The shifts imperfectly follow the delay (autumn) and the advance (spring) of sunrise times.[1–3] All else equal, it is equivalent to a seasonal regulation of the pre-set schedules so that the phase of human activity is bimodal: delayed in winter, advanced in summer because 8am DST is just an alias for 7am ST (standard time), the hour preceding 8am ST. Clock regulations have been long time under public scrutiny[4] due to the acute impact that the stroke of one hour brings in health and societal issues, specifically in traffic accidents[5, 6] and myocardial infarctions[7], and of the disruption of the seasonal adjustement[8]. Some review studies have also put emphasis on the chronic circadian misalignment that the practice might bring[9, 10] and have flagged the practice as “artificial”. With tiny, unspecific energy savings as the only potential benefit of seasonal clocks, many physiologists and some medical societies have built a strong case against the practice and demand decision makers to cancel it[11–15]. The understanding of social preferences for seasonal DST and for permanent DST has also become recently a point of interest in this field. Surveys usually show a majority of respondents pro canceling the biannual change of clocks, which is celebrated by those who sponsor the abolishing of clock regulations. But polls also show a majority pro setting DST permanently, in bold disagreement with the “science-informed” positions that favor permanent ST[16–18]. During the summer of 2018 the European Commission launched a web-based, self-reported, public consultation on summertime arrangements [Daylight Saving Time, DST] in their *Have your say* portal. The public consultation targeted one and every of the then 500 million EU residents. The final report counts *N* = 4 540 554 million replies to the consultation from citizens in the then 28 Member States[19]. From an aprioristic point of view the public consultation was prone to confounding, uncontrolled factors due to the lack of sample design. Respondents were not even questioned about their age, sex, work status, marital status or region. Apparently, little relevant information can be extracted from the results of the consultation other than the bulk descriptive analysis performed by the European Commission and based on the shares of respondents. As an example of these uncontrolled factors it was noted that 70% of the replies came from the same Member State, Germany, which accounts only for 16% of the population. In contrast, this study performs a posterior analysis of the results that unveils the presence of factors that impacted the consultation in a prescribed, controlled way. We search for the explanatory variables the best predict the shares of population against the regulations. If uncontrolled factors dominated the consultation, then no such variable should exist. Else, the existence of such variable is evidence of a strong, controlled factor. The study focuses on one single question of the consultation: question 2, the individual preference *to keep or to abolish the current arrangement?* This one and only question that provides a feedback on the *statu quo* under which the respondent has been living in the recent past. In this view, the public consultation was a natural experiment in which those against or pro the current arrangements were called to have their say. Because of that, and given the size of the experiment, our study has general implication for the current understanding of the issues associated with clock regulations. The current prevailing point of view associates DST regulations with longitude, time zone, and position in time zone.[10, 20] However, the rationale behind the current regulations is linked to the seasonal changes in the cycle of dark and light, and focus on “the alignment of works hours closer to the sunrise and greater use of outdoor recreation in the [spring and summer] evening”.[1–3, 21–23]. In this point of view, clock regulations are impacted by latitude only because latitude is the one and only factor that drives seasonal variability. Previous research studies have noted the impact of latitude and seasons in the sleep-wake cycle on narrow ranges of latitude.[24, 25] For this issue, the public consultation is a dreamed natural experiment, since it targeted population living from 35° latitude to 65° latitude. This study shows that the shares of population against the current regulations correlated (1) with the winter sunrise time, and (2) with the distance of the human starting point of activity to the winter sunrise. Higher latitude —then more delayed winter sunrise time— and earlier starting points relative to the winter sunrise resulted in larger shares of population against the current regulations. These results are in line with the aims and goals of the clock regulations, and with the role of light and the photoreceptive mechanisms in regulating human activity[26] and challenge the current calls to cancel the clock regulations globally. Three major strengths of the study must be highlighted: first the large number of countries involved — from 25 to 17, depending of the analysis— which brings the association between explanatory variables (latitude and human activity) and the outcome (preferences to abolish DST) statistically significant. Second, the large sample size of the public consultation: the median value of the distribution of turn-outs per Member State was 0.367% of the population, much larger than a standard poll or survey. Third, as noted before, the wide range of latitudes that allocates the different Member States which brings vividly different seasonal cycle of light and dark, under the same clock regulations, and give support to the correlation. Previous studies on the association between DST preferences and latitude were limited to one country (Norway[27]) and one province (Alberta, Canada[17]), both with a narrower range of latitude. ## 2. METHODS ### 2.1. The 2018 public consultation on summertime arrangements From 4 July 2018 to 16 August 2018 the European Commission launched a web-based public consultation on summertime arrangements in their *Have your say* portal. The final report counts *N* = 4 540 554 million replies to the consultation from citizens in the then 28 Member States[19]. A clear majority, 84% of respondents, demanded the cancel of DST. A fortnight after the consultation ended, the Juncker Commission launched the proposal to cancel of DST. However Member States could not agree on which time should survive, a most complicated decision to which the consultation gave no clear preference: 56% of respondents preferred summer time, 32%, winter time. Public consultations are not surveys due to the lack of sample extraction design. The European Commission acknowledges that the consultation on time arrangements had “a self-selection bias of the respondents towards the views of those who choose to respond to the consultation”1. Unlike polls, where respondents are urged to answer a question, the public consultation was a self-reported survey where some people choose to respond the consultation. Our study focus only on question 2: the individual preference *to keep or aboslish the current arrangement?* Question 2 was retrospective and provided a feedback on the clock regulations upon which the individuals had lived in the past decades. On the contrary question 5 (the individual preference for permanent winter time or for permanent summer time) is prospective. Replies to this question express, for the majority of the population, preferences for hypothetical, unknown scenarios. These personal preferences may or may not sustain in time if they materialize. More importantly question 2 segmented the respondents in the two categories that matters for the discussion: those against and those pro the current regulations. We analyzed these two population groups separately. We do not test those against DST versus those pro DST. We do not test the net balance of those against and pro DST. We test those against DST irrespective of how many individuals showed up in the public consultation pro DST. And we test those pro DST irrespective of how many individuals showed up in the public consultation against DST. Either group of population is isolated. We took the data from Table 8 (Appendix B.1.2, page 41) in the final report, which lists the absolute numbers of answers by citizens to Question 2 in the consultation: *to keep or abolish the current arrangement?* We also retrieved the population numbers *P* for 2018 by Member State from the Eurostat database (demo_pjanind.tsv) and, from the World Bank database, the fraction of population *I* using the Internet in 2018. While the public consultation intended to target the entire population *P*, its web-based design implied that only the fraction *I* × *P* was effectively targeted. We will make use of this population basis in our study. Since, within the European Union, *I* has a narrow range from 65% to 97%, differences between the choice *P* × *I* and *P* are slight and more technical than substantial. By scaling the absolute respondents to Q2 with the target population we obtain the *shares of (target) population* against the current arrangement *C* by Member State and the *shares of (target) population* in favor of the current regulations *K*, as per the public consultation. We note that the European Commission only reported the *shares of answers* or *shares of respondents* against and in favor of the current regulation in Figure 9 (page 14, Member States ordered in decreasing values of shares against) and in Figure 28 (page 35, Member States ordered in decreasing values of latitude). The Commission also reported the participation rate by Member State: number of respondents divided by population, see Table 2 (pages 8 and 9). A posterior analysis of the shares of target population allows to identify which stimuli prompted the population to take the consultation, see section 4. The quartiles of the distributions of shares were: *Q*1 = 0.204%, *Q*2 = 0.406%, *Q*3 = 0.621% (against), and *Q*1 = 0.030%, *Q*2 = 0.069%, *Q*3 = 0.143% (keeping). The interquartile range was IQR = *Q*3 − *Q*1 = 0.416% (against) and IQR = 0.112% (keeping). While these sizes are small relative to the total population of a country, and small to assess the net balance of the two choices, we note that these sizes are much larger than the sample size of a standard poll, survey or self-reported, web-based studies. As a comparison Kantermann *et al*. [8], a golden-standard reference in DST studies, is a web-based, self-reported study that collected 55 000 replies in Germany, or ∼ 0.09% of the population with access to the Internet in 2007. We computed the distance to the third quartile in units of the IQR. Austria, Germany and Luxembourg showed distances larger than 1.5 in either distribution; Cyprus got a distance larger than 1.5 in the keeping distribution. We note that 1.5 is a common thresohld for outlier identification in descriptive statistics. These outliers were removed from the bivariate linear analyses, see Section 4.1 for the implications. Alternatively, median quantile regressions[28] in the full dataset yield similar results. We stress that the removal of these outliers in no way undermines the choice of these persons. The removal of outliers just allows to characterize bivariate regressions in the remaining dataset. Table S2 lists the population numbers *P*, shares of internet users *I*, the replies to question 2 in the Member States that took part in the consultation, and the distance of the outliers to the third quartile. ### 2.2. Clock regulations and geography In the prevailing point of view DST transitions are associated with time zone[9, 10], and position in a time zone[20]. The latter can be expressed by the time offset *τ* : ![Formula][1] Here *λ* is the longitude, Ω = 15 ° h−1 = 0.262 rad h−1 is Earth’s angular rotation speed and *Z* is the time zone offset: a whole number of hours ahead or behind Universal Time (UT). Clock regulations alter *Z* on transition dates. As an example for the Central European Time zone *Z* = 1 h, while for the Central European Summer Time zone *Z* = 2 h. The time offset *τ* sets the difference between solar time (solar noon) and clock time (midday). With solar noon occurring every 24 h day after day, irrespective of season, and midday occurring every 24 h —we assume no clock regulations here— time offset is not linked to seasons. The rationale behind clock regulations is linked the variability of sunrise times through seasons, which is dominated by latitude *ϕ*. Conveniently the spread of sunrise time in the entire year is related to the seasonal factor *F* (*ϕ*): ![Formula][2] where *ε* = 23.5° is Earth’s axial tilt. Figure S1 in the Supplemental Information shows a scatter plot of *F* (*ϕ*). The factor is defined so that the winter sunrise time (SRW) and summer sunrise time (SRS) are[29]: ![Formula][3] ![Formula][4] where *F* = 6 h is the year-round Equatorial sunrise time. Likewise sunset times go as 18 h ∓ *F* (*ϕ*). It must be noted that all times here are mean solar times; clock time values are obtained by adding the time offset in Eq. (1). With *F* (*ϕ*) a small and slow varying function at low latitudes (see Figure S1 in Supplemental Information) the stroke of one hour set by the clock regulations finds little support in the Tropic. Likewise with *F* (*ϕ*) high and steepest it would find little support in the Polar region, because the varying seasons brings a span of sunrise time much larger than one hour. At intermediate latitudes the stroke of one hour is balanced with better alignment of human starting times with sunrise time, smaller exposition to the damaging central hours of the day during the summer season, and with an improvement on the leisure time during this season[2, 22, 23]. Little is known on how latitude impacts the public preferences for the practice, save for the obvious fact that the practice is seldom used at the Tropics. The European Union is an adequate environment to test this idea because the supranational ensemble extends from 35° (Cyprus) to the Polar Circle (Finland and Sweden), which sets vividly different seasonal cycles of light and dark: sunrise time can span as few as less than three hours (Cyprus) to more than seven hours (Finland). Table S1 in the Supplemental Information lists European Union Member States and United Kingdom in increasing values of latitude, with time offset and the seasonal factor. Figure S2 in Supplemental Information shows an Albers projection of Europe and visualizes longitude and latitude. Geographical data were retrieved from the GeoNames database. We searched for populated places with more than 500 inhabitants. Population weighted low quartile, median and high quartiles for the distribution of latitude and longitude were determined. We also determined the distribution of winter sunrise times per country and their population weighted quartiles. Morning social times, related to the onset of human activity after the nighttime rest, are synchronized to the winter sunrise time over wide ranges of latitude.[29, 30] The distance between social time *t* (local time) and SRW can be expressed from Eq. (1) and Eq. (3) as: ![Formula][5] where *t**s* is the corresponding mean solar time. Positive *t**w* means a starting point in the photoperiod during the entire year; while negative *t**w* means starting point in challenging dark condition, see later section 4. Bright light is known to have an impact in human physiology and in the human circadian system.[26, 31] We retrieved social times from Time Use Surveys. We focus on the daily rhythms available at the Harmonized European Time Use Survey (HETUS)[32], a collection of large scale time use surveys which were carried out in two rounds (2000 and 2010). Nineteen Member States have reported time use statistics to Eurostat in either of the two rounds. Two of them, Germany and Luxembourg, were removed as outliers. Therefore the analysis consists of 17 Member States. We focus on the two activities of interest in connection to the seasonal regulations of time: sleep and other personal care (code AC0A), and work as a primary and secondary activity, including the transfer from home to work, (code AC1_2_TR). The first activity largely reproduces the sleep/wake cycle in the participant state, with a minor contribution of other personal care (washing, clothing and the like). The second activity refers to the labor cycle. Both activities are linked one way or the other to the cycle of light and dark, with which the time regulations are associated. HETUS reports the participation rate in an activity by the local time of the day in 144 distinct, 10-minute time slots. The sum of all participation rates is the fraction of time spent in the activity. The phase of the daily rhythm can be assessed by the midpoint (centroid) of the activity, see Figure S3a in Supplemental Information: the time slot when half the daily time spent has been consumed and half remains. For this purpose the daily rhythm is started when the participation rate is minimal: 04:00 (work) and 19:00 (sleep and other personal cares). We also assessed the starting point of work (or work onset) and the end point of the sleep and other personal cares (or sleep offset) by scaling the daily rhythm with the daily maximum participation rate and retrieving the time slot when the scaled participation rate is larger than one half (work), or smaller than one half (sleep). See figure S3b in Supplemental Information. In addition to the concerns and limitations that the study design of the public consulation presents for our analysis we also mention a further limitation. The two activities here analyzed refer to distinct population structures. The work activity is filled only by the fraction of survey respondents that do work, while the sleep and other cares activity is filled by every individual taken part in the survey. In contrast, it is not possible to extract the responses given by employees only in the 2018 public consultation. ## 3. RESULTS ### 3.1. How the public consultation preferences associate with the geographical predictors We first show the association between the geographical predictors (details in Section 2.2) and the results of the 2018 public consultation on DST (outcome, details in Section 2.1). We consider the two outcomes to questions 2: the shares of respondents willing to cancel DST, hereafter *C*, by Member State; and the shares of respondents willing to keep DST by Member State, hereafter *K*. As noted before, Germany, Austria and Luxembourg were identified as outliers and removed from the following analysis, see section 2.1 for details. For the keep choice Cyprus shares were also qualified as an outlier and were removed from the bivariate analysis. In the end, the sample size was 960 199 replies (median shares 0.390%) in 25 member states (against) and 174 489 (median shares 0.047%) replies in 24 member states (keep). On the other hand, we consider four predictors. First we test the longitude *λ* and the time offset *τ*, see Eq. (1)—the distance from solar noon to clock midday. Neither of these two predictors sets seasonal variability and therefore neither should set preferences for neither outcome. The null hypothesis “the shares of respondents pro canceling *C* or pro keeping *K* DST by Member State do not depend on the longitude or the time offset” sustained (*p* = 0.173; *p* = 0.228; *p* = 0.570; *p* = 0.894) at the standard level of significance (*α* = 0.05), see Table S4 and Figure S4 in the Supplemental Information. We then considered the latitude *ϕ* as a predictor, and found that that the null hypothesis sustains for the abolish outcome (*p* = 0.081; *R*2 = 0.127; *N* = 25) and for the keep outcome (*p* = 0.420; *R*2 = 0.030; *N* = 24). Finally, we tested the SRW, see Eq. (3). We found that the null hypothesis does not sustain (*p* = 0.036, *R*2 = 0.177, *N* = 25) for the abolish choice. Preferences for canceling DST increased with delayed sunrise times and larger seasonal variability, in line with previous expectations[23]. On the contrary the null hypothesis sustained (*p* = 0.420) when the keep choice was tested. Figure 1A shows a scatter plot of *C* versus the winter sunrise time with horizontal lines signaling the low and high quartiles of the distribution by Member State and Figure 1B does the same for the shares *K*. The vertical axes in either panel expand the same range, which allows to perceive the differences between *C* and *K* in size and in their correlation with SRW. Figure 1 in Supplemental Information shows panel B in a magnified scale. In addition, Figure S6 in Supplemental Information shows the association between *C/K* and SRW, which brought larger *C/K* with higher latitude and SWR (*R*2 = 0.34; *p* = 0.003; *N* = 28). ![Figure 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/08/13/2024.03.01.24303549/F1.medium.gif) [Figure 1.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/F1) Figure 1. **A**. The association between the winter sunrise time SRW = *F* +*F* (*ϕ*) (solar time) and the shares of population willing to abolish DST as per the 2018 public consultation *C*. The inset shows the distribution of shares scaled by the interquartile range to identify and quantify the outliers Austria, Germany and Luxembourg, not shown in the main picture and excluded from the regression (see section 4). The low quartile was *Q*1 = 0.204% and the high quartile, *Q*3 = 0.620% (see the boxplot on the left side of the plot). The solid line shows the point estimates of the prediction; the dashed lines bound the 95% prediction interval. The regression is statistically significant (*p* = 0.03) at the standard level of significance (*α* = 0.05), see Table S4 in the Supplemental Information. Notice that the span of the predictor is larger than one hour: the standard unit of social time and the size of the stroke brought by DST of clocks. Horizontal bars highlight population weighted low and high *F* (*ϕ*) quartiles (France and Spain show the span of the European areas only). Labels show iso-3166 alpha-2 codes. **B**. The same picture but for the shares of population willing to keep regulations as per the 2018 public consultation. The association does not bring statistically significant correlations, see Table S4 in Supplemental Information. For the shares *K, Q*1 = 0.030% and *Q*3 = 0.143% (see the boxplot on the left side of the plot). Panels a A and B have the same vertical scale for the sake of comparison; Figure S5 in Supplemental Information shows panel B in a narrower scale. Labels in increasing values of latitude: CY Cyprus; MT Malta; GR Greece; PT Portugal; ES Spain; BG Bulgaria; IT Italy; HR Croatia; RO Romania; SI Slovenia; HU Hungary; FR France; AT Austria; SK Slovakia; LU Luxembourg; CZ Czech Republic; BE Belgium; DE Germany; PL Poland; GB United Kingdom; NL Netherlands; IE Ireland; LT Lithuania; DK Denmark; LV Latvia; SE Sweden; EE Estonia; FI Finland. We bring the attention to the pivotal role played by physical geography in the statistical significance in the *C* versus winter sunrise time association. The range of the predictor extends for two hours, larger than the size of the DST shift. ### 3.2. How the public consultation preferences associate with human activity related predictors To understand the interplay between the shares of population against DST, environmental factors and human physiology we now analyze their association with the human activity. The data set is here reduced to the *N* = 18 participants in HETUS that also participated in the 2018 public consultation. On of them (Germany) was flagged as on outlier. The number of replies in the sample of *N* = 17 Member States adds up 813 564, with median shares per Member State 0.390%. We did not find any statistically significant association between *C* and the sleep/wake cycle or the work cycle when their parameters are measured by a distance to noon (either clock time or mean solar time). This is in line with our previous expectations because clock time and mean solar time are not impacted by seasons or latitude. On the contrary, we did find a statistically significant association when the distances of the midpoints of the sleep/wake cycle and the work cycle to the winter sunrise time were tested. For the association between the sleep/wake cycle and *C* we found *p* = 0.009; *R*2 = 0.38; while for the labor cycle we found *p* = 0.002; *R*2 = 0.47. Either of the two activities finds larger shares of population willing to cancel DST with earlier midpoints, shown by the negative value of the bivariate slope; see Table S3 in Supplemental Information. In line with the goals of DST, which aims to regulate starting times, we analyzed the association between *C* and the distance of morning time marks —sleep offset and work onset— to SRW. For the sleep offset we found *p* = 0.003; *R*2 = 0.44. For the work onset we found *p* = 5 × 10−5; *R*2 = 0.68. Notably the shares *C* correlate the more against the morning time marks for both activities. These results improve those obtained when the shares *C* were tested against the latitude (*p* = 0.018; *R*2 = 0.32) and the winter sunrise time (*p* = 0.008; *R*2 = 0.38) in this subset of *N* = 17 Member States. Figure 2 shows in four panels the association between the morning timemarks (left) and midpoints (right) with the shares of target population willing to cancel DST as per the public consultation. The top two panels refer to the sleep and other cares activity: sleep offset (left) and sleep midpoint (right). The bottom two panels refer to work onset (left) and work midpoint (right). The solid line shows the regression and the dashed lines bound the 95% prediction interval associated with the regression. ![Figure 2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/08/13/2024.03.01.24303549/F2.medium.gif) [Figure 2.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/F2) Figure 2. Scatter plots of the shares of target population willing to abolish DST as per the 2018 European Commission public consultation (outcome) vs four predictors related to human activity as per HETUS. The left column shows morning time marks: sleep and other personal care offset (top) and work onset (bottom). The right column shows noon time marks: sleep and other personal cares midpoint (top), work midpoint (bottom). Time marks are converted into distance to the winter sunrise (SRW), see Eq. (5). Circles refer to round 2 (year 2010) of HETUS, and squares to round 1 (year 2000). The null hypothesis “the predictor does not explain the outcome” does not sustain at the standard level of significance for every test, see Table S5 in Supplemental Information. Solid line shows the results of the linear regression; dashed lines bound the 95% prediction interval of the regression. Table S1 lists the data values that generated this plot. Horizontal bars highlight population weighted low and high quartiles (France and Spain show the span of the European areas only). ## 4. DISCUSSION ### 4.1. Statistical significance of the correlations and rationale behind The 2018 public consultation was a natural experiment driven by a self-reported study where “the respondents had to give very little information about themselves.”[19, Section 3] The results were only disaggregated by Member State and the final report on the public consultation reads: “caution needs to be taken for the interpretation of the data in terms of causality and one needs to remember that the results are not [statistically] representative. This is why *only [univariate] descriptive statistics* are presented in the report.”[19, Section 3] The final reports also highlights that “even with the caveats in terms of biases and non-representation, the results can be analysed by taking into account an important differentiating factor between the Member States, namely their location and their exposure to more or less light [driven by the latitude].” Indeed the Appendix A3 in the final report considered three latitudinal clusters: above the 53rd parallel (7 Member States), below the 46th parallel (9) and the intermediate range (12). However latitude is only employed to report in Figure 28 (Appendix A.4.2, page 35) the shares of respondents to Question 2 with Member States ordered by decreasing latitude of the capital city. Our study addresses specifically these challenges and refines the statistical analysis. We test the shares of (target) population —not the shares of respondents— the most sensitive metric in a natural experiment. Aside from the evident bulk differences in the shares of respondents and in the shares of population against and pro the current regulations that drew the attention in 2018, Figure 1 and Table S4 show that the shares of population against *C* and pro *K* the current regulations found different stimuli to show up at the public consultation. While the keep choice finds no relevant correlation with any metric, in line with the uncontrolled factors that impacted the public consultation, the abolish choice finds correlations with *ϕ* (*R*2 = 0.13; *p* = 0.08), with SRW (*R*2 = 0.18; *p* = 0.01), and with *t**w* (*R*2 = 0.68; *p* = 5 × 10−5). One kind of stimuli brought larger shares of population willing to cancel the regulations with increasing latitude, with delayed winter sunrise time and with larger shares of early risers (those who start their activity before the winter sunrise time). This stimulus was able to overplay the myriad of confounding factors that might have also impacted the distribution of shares. We note the bare meaning of *p* = 5 × 10−5: if the distribution of shares *C* is shuffled —simulating strong uncontrolled confounding factors only— and the shuffled distribution is tested against the predictor, there are 5 odds in 100 000 tests of obtaining a greater correlation. It is very unlikely that the correlations in figure 2 arose by chance. In obtaining this result the large sample size per Member State of the shares *C* (median value ∼ 0.390% of the target population) must have played a significant role: it was able to retain significant unbiased information of the underlying population even in an uncontrolled, natural experiment. We suggest that this primordial stimulus is linked to the primordial rationale of the clock regulations: “the alignment of works hours closer to the sunrise”:[2, 21, 22, 33] those who lose synchronization with SRW and advance their phase relative to it, find little outcomes in further advances of the activity in spring, driven by the clock regulations. The stimulus can also be linked to human physiology:[3, 23, 34] the early risers find challenging dark conditions in the winter morning; they may find dark challenging conditions after DST spring forward shift, and before the autumn shift, either case depending on the transition dates; finally, they also find bright challenging conditions during the summer evening, which jeopardizes their bedtimes. We must note that SRW and the summer sunset (SSS) are always separated by twelve hours. Early risers have a shorter gap between their onset and SSS, which is further shortened by clock regulations, jeopardizing bedtimes. This rationale is irrespective of latitude; increasing latitude simply promotes larger shares of early risers as a result of increasingly delayed winter sunrise times. On the contrary, the late risers —those persons with start times in line with the winter sunrise— seldom find these issues and show little discomfort with the practice: with an onset time delayed by the delay in WRS they find benefits in advancing their onset times in spring-summer. This group of population did not find the stimulus to show up at the consultation. A big, unavoidable limitation of our study is the coarse disaggregation by Member State. This issue questions whether the reported mean shares *C* of a given Member State would sustain for every region of the Member State or, else, if a distribution of shares *C* by smaller statistical units might have been observed. The Commission attributed little significance to this issue because the respondents were not even asked for region. In the final report, this issue is indirectly addressed when the differences in the exposure to more or less light —driven by the latitude— are noted as important differentiating factors. The sentence is meant at the level of Member States because conditions in Southern States are widely different from conditions in Northern States. However this range of exposure to more or less light is less significant at a finer level in every Member State except perhaps in Great Britain, Sweden and Finland, where differences between their Southern regions and their High North regions are still significant due to the divergence of *F* (*ϕ*) at their range of latitudes. We note in Figure 1 and Figure 2 interquartile ranges of the predictor at State level, see the horizontal bars. In the end, the analysis by Member State does not discard our results. On the contrary, our results urge to analyze the relationship between clock regulations and latitude at a finer level of disaggregation. Finding correlations between social metrics by Member State and geographical factors is not rare in the European Union and elsewhere. As an example, shares of population using internet *I* increases with *ϕ* or with SRW (*R*2 = 0.27; *p* = 0.004; *N* = 28). Also the Gross Domestic Product per capita in 2018 increases with *ϕ* or with SRW (*R*2 = 0.34; *p* = 0.002; *N* = 25) once low populated countries (outliers) are removed from the association. The important thing to note in the association between the preferences for canceling clock regulations and *ϕ*, or SRW is the direct link between the explanatory variable and the outcome, and a rationale behind the correlation. In addition, the association between the shares against the clock regulations and *t**w* highlights only the ambient light conditions that punctuate the onset of human activity after the nighttime rest. A counterexample of confounding factor with relevant impact in the distribution of shares is shown by the shares of Austria, Germany and Luxembourg where a clear, distinct, unidentified, strong confounding factor brought larger shares of population to show up at the public consultation. We note that the three Member States are not outliers for the metric *C/K* (see Figure S6 in Supplemental Information), meaning that this confounding factor was not biased towards one of the population groups, but played globally. As a plausible hypothesis we put forward the following argument. In order to show up at the public consultation the general population first needed to know about it from public advertisement launched by the European Commission and, above all, from the impact of the consultation in mass media. We note that Austria and Germany are the two only member states where the German language is the official language; while in Luxembourg, the German language is spoken by significant shares for the population. It might have been the case that German language media put more attention on the public consultation than elsewhere. Larger shares of audiences brought larger shares of respondents compared with other Member States. ### 4.2. Further implications of the study Our findings are counter the prevailing point of view described by position statements issued by medical associations[11, 15, 35]. It is showing that in order to build up shares of population against the “annoying” DST regulations that plays against the circadian (twentyfour hours) rhythm, shares of early risers must be built up as well. They would face “annoying” dark mornings, that plays against the photoreceptive mechanisms in human physiology.[26] The circadian golden standard of year-round pre-set schedules with start times aligned with the winter sunrise seems to be unstable in some societies. Indeed, this golden standard must have been present at the beginning of the 20th century, and it was soon deprecated in United Kingdom, urban areas of United States, Ireland and Portugal, by giving birth to the biannual changing of the clocks. Eventually, by advancing clocks in spring and delaying it back in autumn, these societies prevented the growing use of artificial light from incorporating more human activity into the dark winter morning[36]. As a result, the biannual clock change kept the number or early risers at bay; the shares of population against the regulations did not grow either; and the regulations probed stable. Our results also challenge the frequent argument that the seasonal clock regulations bring a circadian misalignment during the DST period because the social clock is “artificially” out of sync with the sun clock[9]. This claim is supported by research studies that associate position in a time zone with hazards.[20] If DST brought a circadian misalignment due to the misalignment between the social clock and the sun clock, then the shares of population against DST in the public consultation and the latitude should not correlate, because the misalignment does not depend on the latitude. The misunderstanding arises because the misalignment is associated with clock time and with mean solar time (noon). However, human preferences and practical decisions are connected with sunrise, and therefore with latitude and seasons[29]. As long as the clock regulations do not bring the starting points ahead of the sunrise, and the bedtimes ahead of the sunset, the alleged misalignment is irrelevant for practical decisions and, very likely, a matter of personal preferences.2 Our results may also help to understand why many polls display larger preferences for permanent DST, a choice that bothers the circadian community[16, 17]. More than expressing a science-informed preference for this setting, the polls might just be expressing that newer generations still love their current summer earlier start times that allow a better harvesting of their leisure time. They do not see late start times and late end times during the summer season. Because the shares of population pro DST did not show up consistently in the 2018 public consultation, our methodology cannot assess whether above some circle of latitude the majority is against the current regulations. Nonetheless, the findings of this manuscript, supported by such large sample size, have a substantial impact in the current wave of discussion around summertime arrangements and, specifically, challenges the one-size-fits-all solutions that usually show up. As an example, we put forward the commitment of the European Commission to a general policy: either every Member State must practice summertime arrangements (the *statu quo* implemented through Directive 2000/48/EC) or they all must abandon the practice. In sharp contrast, our methodology find a significant stratification of preferences with latitude and with human activity. Therefore our study suggests that a science evidence-based policy should allow Member States to opt-out from the current setting, and should implement binding transition dates only for those Member States who wanted to continue the seasonal practice. Chile and Australia align with this choice and clock regulations stratify with latitude in these countries. Compared to the European Union, the contiguous USA happen to experience more similar seasonal cycles of light and dark, with perhaps the exception of Florida, close to the Tropic of Cancer. We do not know of a public consultation in USA similar to the public consultation launched by the European Commission. Yet, the US starting point of human activity is in line with the winter sunrise[29], as observed in United Kingdom or Italy. Therefore we would expect the US general population would not strongly oppose seasonal DST in self-reported surveys. This would also be in line with the long standing record of the practice (some one hundred years in the US urban areas). On the contrary, the Russian starting point of human activity is in line with the Scandinavian and Baltic starting points as deduced from reported values of school starting times[37]. As a result, year-round DST failed to sustain in Russia after 2015, when year-round ST was adopted. While this choice was celebrated by the circadian community we bring the attention to the fact that this is a consequence of having previously advanced the starting point of human activity, due to the challenging conditions associated with the Russian winter photoperiod, and may not necessarily apply elsewhere. ## 5. CONCLUSION This study unveils evidence that the 2018 Public Consultation by the European Commission, a natural experiment, got replies in line with human physiology and with the aims and goals of the clock regulations. Those with earlier onset of their activity relative to their winter sunrise time were proner to say “no” to the regulations. The bi-annual change of the clocks is abrupt, annoying and hazardous, but societies seldom practice them for the fun of it, or only for economic reasons. They happen to manifest a simple, physiological adaptation to seasons once pre-set schedules were established, and when the environment demands it. That is within a range of intermediate latitudes where *F* (*ϕ*) ranges from 1 h and 2 h (30° to 50° in latitude) so that DST effectively binds the starting point of human activity to the sunrise (see the annotated box in Supplemental Information, Figure S1). The onset of human activity plays a leading role in understanding the preferences for clock regulations. This fact jeopardizes the seek for general solutions of the kind “clock regulations must be canceled” or “clock regulations must be enforced” in wide areas like the European Union. ## Data Availability All data produced in the present work are contained in the manuscript. A data-set is accesible at https://doi.org/10.5281/zenodo.13301201 [https://doi.org/10.5281/zenodo.13301201](https://doi.org/10.5281/zenodo.13301201) ## 6. DATA ACCESSIBILITY STATEMENT A dataset for this manuscript is available at zenodo under doi:10.5281/zenodo.13301201. The most recent version of the dataset is available under doi: 10.5281/zenodo.13301200. Results from the public consultation are publicly available in tabular format[19]. We only made use of the column “Abolishing the switching for all EU-MS Citizens” in Table 8, Section B.1.2, page 41. We discarded the 8758 replies not assigned to any Member State. HETUS data are publicly available at Eurostat[32] for all Member States that participated in the survey except Sweden. We took Sweden results from an earlier study[29]. In the HETUS database we took the participation rate in the main activity (wide groups) by sex and time of the day, file tus_00startime. European population numbers for 2018 were extracted from Eurostat’s Demography, population stock and balance database: population (national level), population on 1 January by age and sex, file demo_pjan. Individuals using the Internet as percent of population in 2018 were retrivied from the World Bank indicator IT.NET.USER.ZS available at [https://data.worldbank.org/indicator/IT.NET.USER.ZS](https://data.worldbank.org/indicator/IT.NET.USER.ZS). This indicator is provided to World Bank by the International Telecommunication Union (ITU). Gross Domestic Product per capita in 2018 (current USD) was retrieved from the World Bank indicator NY.GDP.PCAP.CD available at [https://data.worldbank.org/indicator/NY.GDP.PCAP.CD](https://data.worldbank.org/indicator/NY.GDP.PCAP.CD). Geographical coordinates (latitude, longitude, time zone offset) were extracted from GeoNames database ([https://www.geonames.org/](https://www.geonames.org/)). We retrieved the file cities500.zip, and produced population quartile values of latitude, longitude and winter sunrise time for every country in the public consultation. We exclude population places below 500 inhabitants and places flagged with the PPLX feature code (section of populated place). The datafile was downloaded on 22 April 2024. ## DISCLAIMER Along this manuscript United Kingdom is listed as a EU Member State since it participated as such in the public consultation of 2018 and in HETUS rounds. United Kingdom ceased to be a EU Member State on February 1, 2020. Removing the British results from the field does not alter the findings of our study in a substantial way. ## SUPPLEMENTAL INFORMATION Supplemental information is attached to this document as an appendix. ## Supplemental information ### Appendix A: Tabular material This section is intended to be included in a dedicated Supplemental Information file, should the manuscript be accepted and published. It can be tailored to comply with the journal’s Guidelines. View this table: [Table S1.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/T1) Table S1. The European Union Member States at the time of the 2018 public consultation with geographical coordinates: latitude *ϕ*, longitude *λ*, time offset —the difference between noon and midday, *Z* − *λ/*Ω, where *Z* is the time zone offset and Ω is Earth’s rotation speed— and the seasonall factor *F* (*ϕ*), see Equation (2). View this table: [Table S2.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/T2) Table S2. The results to question 2 in the 2018 public consultation launched by the European Commission with population numbers *P* and shares of population with access to the Internet *I*. Member States are listed in increasing values of latitude. Each answer to the question is given in absolute numbers and in shares of the target population *P* × *I*. The quartiles of the distributions of shares are: *Q*1 = 0.204%, *Q*2 = 0.406%, *Q*3 = 0.620% (against), and *Q*1 = 0.030%, *Q*2 = 0.069%, *Q*3 = 0.142% (keeping). The distance to the third quartile in units of the interquartile distance (*Q*3 − *Q*1) is shown in parenthesis when this metric is larger than 1.5, a common criterion for outlier identification. View this table: [Table S3.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/T3) Table S3. The time marks associated with the sleep and other personal cares and the work cycles as per HETUS. Countries are listed in increasing values of latitude. The third left-most column list the round of HETUS: round 1 was taken around 2000; round 2 around 2010. Round 3 is currently in progress. Figure S3 shows the methodology to asses midpoints and offset/onset time marks. View this table: [Table S4.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/T4) Table S4. The association between the geographical data (predictors) and the shares of target population willing to abolish *C* or willing to keep *K* the current time arrangements as per the 2018 public consultation launched by the European Commission (outcome). The null hypothesis fails to sustain for seasonal predictors and the abolish choice (see Figure 1 in the main text) and sustains for non-seasonal predictors (see Figure S4 in the Supplemental Information) and for the keep choice (see Figure 1). The unbiased standard error (SE) is scaled by the sample average value of the either outcome (0.394%, against, and 0.074%, keeping). The null hypothesis of normally distributed residuals sustained in every regression at the standard level of significance. The slopes are given in basis points per degree (latitude) and basis points per hour (SRW). The positive value shows larger shares of target population willing to abolish the current time regulations with higher latitude and higher SRW. The sample size was 960 199 replies to the public consultation from 25 member states for the abolish choice, and 174 489 replies from 24 member states for the keep choice. View this table: [Table S5.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/T5) Table S5. The association between time marks of human activity —measured through two distinct metrics— (predictor, see Table S3) and the shares of target population willing to cancel the time regulations as per the 2018 public consultation (outcome, see Table S2) in *N* = 17 Member States that reported Hetus statistics. Blue ink annotates *p* − values below the standard level of significance (*α* = 0.05), which occurs only when predictors are measured relative to the winter sunrise. The negative slope shows that shares against the current regulations increases with earlier starting points relative to winter sunrise time. Slopes are given in basis points per hour. The unbiased standard error of the fit is scaled by sample average value of the fitted shares *C* (0.390%) and the result expressed as a percentage. The null hypothesis of normally distributed residuals sustains in every regression at the standard level of significance. For comparison the association with geographical predictions in this subset is also shown. ### Appendix B: Additional figures This section is intended to be included in a dedicated Supplemental Information file, should the manuscript be accepted and published. It can be tailored to comply with the journal’s Guidelines. ![Figure S1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/08/13/2024.03.01.24303549/F3.medium.gif) [Figure S1.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/F3) Figure S1. The evolution of the seasonal factor *F* as a function of latitude *ϕ*. The seasonal factor is defined in Eq. (2) so that it equals to zero at the Equator. The seasonal factor gives the advance or delay of the sunrise and sunset times at the solstices relative to the Equatorial standard. The right axis shows the corresponding winter photoperiod (and the summer scotoperiod) 12 h − 2*F* (*ϕ*). Data points annotate the Member States. The seasonal factor was modeled for a point-sized Sun and a refractionless atmosphere. The box expands from 30° to 50° and yields *F* (*ϕ*) from 1 h to 2 h. It is an educated guess for the utility range of seasonal regulations of time. ![Figure S2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/08/13/2024.03.01.24303549/F4.medium.gif) [Figure S2.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/F4) Figure S2. An Albers projection of Europe with the countries participating in the 2018 European Commission public consultation on summertime arrangements (in shade and labeled). Dashed meridians annotate time meridians. Solid meridians annotate the bounduaries of the physical time zones. Circles of latitudes annotate straight values of *F* (*ϕ*), starting at *F* = 1.5 h (southern most) and ending at 3.5 h (northern most) in steps of 0.5 h. Color shades highlight time zones; from west to east, Western European Time (*Z* = 0), Central European Time (*Z* = 1 h) and Eastern European Time (*Z* = 2 h). Vertical bar annotates the quartile range of population weighted latitude at the median population weighted latitude. The black horizontal bar locates the pupulation median latitude and longitude for the country, listed in Table S1. The map was made with Natural Earth, a free vector and raster map data available at [https://www.naturalearthdata.com/](https://www.naturalearthdata.com/). ![Figure S3.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/08/13/2024.03.01.24303549/F5.medium.gif) [Figure S3.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/F5) Figure S3. Determination of the midpoint and work onset. On the left panel the daily rhythm of labor as per HETUS. The area enclosed by the daily rhythm scales with the daily average total work. The shaded areas break even the total area. The hour of the day when this happens is the *midpoint* of the labor cycle. On the right panel, the maximum participation rate is halved. Work onset is determined as the hour of the day when the participation rate first exceeds the maximum daily participation rate. Similar analysis allow to assess the midpoint of the sleep/wake cycle and the sleep offset. ![Figure S4.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/08/13/2024.03.01.24303549/F6.medium.gif) [Figure S4.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/F6) Figure S4. The association between the time offset —difference between noon and midday— and the shares of target population willing to cancel the regulations as per the 2018 public consultation *C*. Germany (3.18%), Austria (2.27%) and Luxembourg (1.38%) are not shown in the picture. The solid line shows the point estimates of the prediction; the dashed lines bound the 95% prediction interval. The null hypothesis “shares *C* do not depend on time offset” sustains (*p* = 0.696 at the standard level of significance (*α* = 0.05), see Table S4 in the Supplemental Information. Labels show iso-3166 alpha-2 codes, see Table S1. ![Figure S5.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/08/13/2024.03.01.24303549/F7.medium.gif) [Figure S5.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/F7) Figure S5. The same as Figure 1B but in a magnified vertical scale to visualize the distribution of shares against the regulations. ![Figure S6.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/08/13/2024.03.01.24303549/F8.medium.gif) [Figure S6.](http://medrxiv.org/content/early/2024/08/13/2024.03.01.24303549/F8) Figure S6. The association between the winter sunrise time WSR and the ratio of respondents against and pro the regulations *C/K*. Note the log scale in the vertical axis intended to appreciate large *C/K* and low *C/K*. The solid line is the quadratic regression (*R*2 = 0.34; *p* = 0.005; *N* = 28); the dash lines highlight the 95% prediction interval of the regression. The log(*C/K*) quartiles were *Q*1 = 1.326; *Q*2 = 1.643; *Q*3 = 1.968, or [3.78, 5.17, 7.15] in the linear scale (see the boxplot on the left side of the plot. Cyprus, Greece and Malta have scores below *Q*1 − 1.5 × IRQ. If they are removed from the regression analysis, then the association is not statistically significant. Labels show iso-3166 alpha-2 codes, see Table S1. ## Footnotes * † jorge.mira{at}usc.es; [https://ror.org/030eybx10](https://ror.org/030eybx10) * We rescaled the population basis and recomputed the correlations. Two figures were added to Supplemental Information. We added a data-set. * 1 In “Public Consultation On EU Summertime Arrangements Report of Results” a working documment by the Commission Staff accompanying the document “Proposal for a Directive of the European Parliament and of the Council discontinuing seasonal changes of time and repealing Directive 2000/84/EC”, document number 52018SC0406, [https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52018SC0406](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A52018SC0406) * 2 Eventually, beyond the Polar Circle and during the Polar night or during the midnight sun period, few can be said about the alignment of the human activity in terms of clock time. * Received March 1, 2024. * Revision received August 13, 2024. * Accepted August 13, 2024. * © 2024, Posted by Cold Spring Harbor Laboratory This pre-print is available under a Creative Commons License (Attribution 4.0 International), CC BY 4.0, as described at [http://creativecommons.org/licenses/by/4.0/](http://creativecommons.org/licenses/by/4.0/) ## References 1. [1]. G. V. Hudson, On seasonal time-adjustment in countries south of lat. 30 degrees, Transactions and Proceedings of the Royal Society of New Zealand 28, 734 (1895). 2. [2]. G. V. Hudson, On seasonal time, Transactions and Proceedings of the New Zealand Royal Society 31, 577 (1898). 3. [3]. J. M. Martín-Olalla and J. Mira, It is time to understand daylight saving time, Sleep 46, zsac309 (2023). 4. [4]. T. H. Monk and S. Folkard, Adjusting to the changes to and from daylight saving time, Nature 261, 688 (1976). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/261688a0&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=934309&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F08%2F13%2F2024.03.01.24303549.atom) 5. [5]. S. Coren, Daylight savings time and traffic accidents, New England Journal of Medicine 334, 924 (1996). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1056/NEJM199604043341416&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=8596592&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F08%2F13%2F2024.03.01.24303549.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1996UD37400028&link_type=ISI) 6. [6]. J. Fritz, T. VoPham, K. P. Wright, and C. Vetter, A chronobiological evaluation of the acute effects of day-light saving time on traffic accident risk, Current Biology 30, 729 (2020). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.cub.2019.12.045&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32008905&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F08%2F13%2F2024.03.01.24303549.atom) 7. [7]. I. Janszky and R. Ljung, Shifts to and from daylight saving time and incidence of myocardial infarction, New England Journal of Medicine 359, 1966 (2008). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1056/NEJMc0807104&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18971502&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F08%2F13%2F2024.03.01.24303549.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000260454500030&link_type=ISI) 8. [8]. T. Kantermann, M. Juda, M. Merrow, and T. Roenneberg, The human circadian clock’s seasonal adjustment is disrupted by daylight saving time, Current Biology 17, 1996 (2007). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.cub.2007.10.025&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=17964164&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F08%2F13%2F2024.03.01.24303549.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000251274300036&link_type=ISI) 9. [9]. K. G. Johnson and B. A. Malow, Daylight saving time: Neurological and neuropsychological implications, Current Sleep Medicine Reports 8, 86 (2022). 10. [10]. T. Roenneberg, E. C. Winnebeck, and E. B. Klerman, Daylight saving time and artificial time zones – a battle between biological and social times, Frontiers in Physiology 10, 944 (2019). 11. [11]. T. Roenneberg, A. Wirz-Justice, D. J. Skene, S. Ancoli-Israel, K. P. Wright, D.-J. Dijk, P. Zee, M. R. Gorman, E. C. Winnebeck, and E. B. Klerman, Why should we abolish daylight saving time?, Journal of Biological Rhythms 34, 227 (2019). 12. [12]. M. M. e Cruz, M. Miyazawa, R. Manfredini, D. Cardinali, J. A. Madrid, R. Reiter, J. F. Araujo, R. Agostinho, and D. Acuña-Castroviejo, Impact of daylight saving time on circadian timing system: An expert statement, European Journal of Internal Medicine 60, 1 (2019). 13. [13]. N. F. Watson, Time to show leadership on the daylight saving time debate., Journal of clinical sleep medicine: JCSM : official publication of the American Academy of Sleep Medicine 15, 815 (2019). 14. [14]. B. A. Malow, It is time to abolish the clock change and adopt permanent standard time in the united states: a sleep research society position statement, Sleep 45, doi:10.1093/sleep/zsac236 (2022). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/sleep/zsac236&link_type=DOI) 15. [15]. M. A. Rishi, J. Y. Cheng, A. R. Strang, K. Sexton-Radek, G. Ganguly, A. Licis, E. E. Flynn-Evans, M. W. Berneking, R. Bhui, J. Creamer,. V. Kundel,. Andrew, R. Spector, O. Olaoye, S. D. Hashmi,. F. Abbasi-Feinberg, Alexandre, R. Abreu, I. Gurubhagavatula,. Vishesh, K. Kapur, D. Kuhlmann, J. Martin, E. Olson,. S. Patil, J. Rowley, A. Shelgikar,. Lynn, M. Trotti,. Emerson, M. Wickwire, S. S. Sullivan, and D. Geffen, Permanent standard time is the optimal choice for health and safety: an american academy of sleep medicine position statement, Journal of Clinical Sleep Medicine 20, 121 (2024). 16. [16]. A. N. Coogan, S. Richardson, and S. Raman, A data-informed perspective on public preferences for retaining or abolishing biannual clock changes, Journal of Biological Rhythms 37, 351 (2022). 17. [17]. M. C. Antle, M. Moshirpour, P. R. Blakely, K. Horsley, C. J. Charlton, and V. Hu, Longitudinal location influences preference for daylight saving time., Journal of biological rhythms 37, 343 (2022). 18. [18]. R. Rubin, Groundswell grows for permanent day-light saving time, but medical societies overwhelmingly support year-round standard time, JAMA doi:10.1001/JAMA.2023.0159 (2023). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/JAMA.2023.0159&link_type=DOI) 19. [19].E. C. D. for Mobility and Transport, Technical assistance with the public consultation on EU summertime arrangements: final report (Publications Office, 2019). 20. [20]. O. Giuntella and F. Mazzonna, Sunset time and the economic effects of social jetlag: evidence from us time zone borders, Journal of Health Economics 65, 210 (2019). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.jhealeco.2019.03.007&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31030116&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F08%2F13%2F2024.03.01.24303549.atom) 21. [21]. W. Willet, The waste of daylight (1907) p. 17. 22. [22]. Y. Harrison, The impact of daylight saving time on sleep and related behaviours, Sleep Medicine Reviews 17, 285 (2013). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.smrv.2012.10.001&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=23477947&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F08%2F13%2F2024.03.01.24303549.atom) 23. [23]. J. M. Martín-Olalla, Comment to “impact of daylight saving time on circadian timing system: An expert statement”, European Journal of Internal Medicine 62C, e18 (2019). 24. [24]. G. Yetish, H. Kaplan, M. Gurven, B. Wood, H. Pontzer, P. R. Manger, C. Wilson, R. McGregor, and J. M. Siegel, Natural sleep and its seasonal variations in three pre-industrial societies., Current Biology 25, 2862 (2015). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.cub.2015.09.046&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26480842&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F08%2F13%2F2024.03.01.24303549.atom) 25. [25]. D. Monsivais, K. Bhattacharya, A. Ghosh, R. I. M. Dunbar, and K. Kaski, Seasonal and geographical impact on human resting periods, Scientific Reports 7, 10717 (2017). 26. [26]. T. Roenneberg, T. Kantermann, M. Juda, C. Vetter, and K. V. Allebrandt, Light and the human circadian clock, Handbook of Experimental Pharmacology 217, 311 (2013). 27. [27]. B. Bjorvatn, S. Pallesen, I. W. Saxvig, and S. Waage, Daylight saving time preferences in norway: Do individual chronotype and home address’ latitude and longitude matter?, Chronobiology International 38, 1449 (2021). 28. [28]. R. Koenker and K. F. Hallock, Quantile regression, Journal of Economic Perspectives 15, 143 (2001). [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1257/jep.15.4.143&link_type=DOI) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000172742200011&link_type=ISI) 29. [29]. J. M. Martín-Olalla, Latitudinal trends in human primary activities: characterizing the winter day as a synchronizer, Scientific Reports 8, 5350 (2018). 30. [30]. J. M. Martín-Olalla, Seasonal synchronization of sleep timing in industrial and pre-industrial societies, Scientific Reports 9, 6772 (2019). 31. [31]. C. A. Czeisler, J. S. Allan, S. H. Strogatz, J. M. Ronda, R. Sánchez, C. D. Ríos, W. O. Freitag, G. S. Richardson, and R. E. Kronauer, Bright light resets the human circadian pacemaker independent of the timing of the sleep-wake cycle, Science 233, 667 (1986). [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Mzoic2NpIjtzOjU6InJlc2lkIjtzOjEyOiIyMzMvNDc2NC82NjciO3M6NDoiYXRvbSI7czo1MDoiL21lZHJ4aXYvZWFybHkvMjAyNC8wOC8xMy8yMDI0LjAzLjAxLjI0MzAzNTQ5LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 32. [32].Eurostat, Harmonised european time use surveys (hetus). rounds 1 and 2 (2010). 33. [33]. O. Pollack, The daylight savings time movement, History Today 31 (1981). 34. [34]. J. M. Martín-Olalla, The long term impact of daylight saving time regulations in daily life at several circles of latitude, Scientific Reports 9, 18466 (2019). 35. [35]. B. A. Malow, Author’s response to “it is time to understand daylight saving time”, Sleep 46, zsac323 (2023). 36. [36]. J. M. Martín-Olalla, A chronobiological evaluation of the risks of canceling daylight saving time, Chronobiology International 39, 1 (2022). 37. [37]. M. F. Borisenkov, T. A. Tserne, A. S. Panev, E. S. Kuznetsova, N. B. Petrova, V. D. Timonin, S. N. Kolomeichuk, I. A. Vinogradova, M. S. Kovyazina, N. A. Khokhlov, A. L. Kosova, and O. N. Kasyanova, Seven-year survey of sleep timing in russian children and adolescents: chronic 1-h forward transition of social clock is associated with increased social jetlag and winter pattern of mood seasonality, Chronobiology International 48, 3 (2017). [1]: /embed/graphic-1.gif [2]: /embed/graphic-2.gif [3]: /embed/graphic-3.gif [4]: /embed/graphic-4.gif [5]: /embed/graphic-5.gif