Perspective: Daylight Saving Time—An Advocacy for a Balanced View and against Fanning Fear
Round 1
Reviewer 1 Report
In the Abstract and Introduction, the work clearly states objectives that are not fulfilled. It not surpasses the level of criticism to other works. In science, before new theories arises, there is a period of discussion between various acceptable theories. The author is pointing out this circumstance and sometimes in a biased manner.
1. Line 76. "Do we need more research?" There is not an answer at the end of the section.
2. Line 113. "the correlation is positive" is erroneous. It is impossible to obtain an overall negative correlation if all the individuals have a positive one. It is more appropriate "the correlation could be positive".
3. Line 115. "Even odds ratios do not provide an intuitive understanding" What is intuitive? It is an universal property that not depends on the culture or academic level of the reader or the ability of the lecturer? Hence, theory of relativity is invalidated if it is not intuitively exposed?
4. Lines 137. "however recently claimed that it was 'biologically large'". That someone says something, it does not prove any kind of crisis in a field. It would be a good idea to gather other works in the same are and recompile their conclusions about the statistical size.
5. Line 198. "In addition to this small and biased sample". Why? The authors have analyzed the power of the analysis or the samples used comparing them with the population?
6. Line 200. "This survey and its results are highly unrepresentative". Therefore, the survey must be rejected? If it is the case, some numbers must precede this conclusion. Saying something is not the same as proving it.
7. Line 231. "...the current state of research suggests that the risk of negative effects associated". This statement is written in the Conclusions sections. It is impossible to conclude this from the previous sections. If there are contradictory researches, some of them shown in the text, why give the authority to one faction?
8. Line 239. "...Therefore, we encourage a discourse, which aims at getting the public on board..." Before the Conclusions, there is not any paragraph discussing about this issue. In the Abstract is written in Line 27 "therefore recommend strategies that could support a discourse aiming". There are not strategies or section talking about them.
Author Response
Reviewer 1
In the Abstract and Introduction, the work clearly states objectives that are not fulfilled. It not surpasses the level of criticism to other works. In science, before new theories arises, there is a period of discussion between various acceptable theories. The author is pointing out this circumstance and sometimes in a biased manner.
We would like to thank the reviewer for his/her time and efforts and the helpful suggestions and critical remarks. We answer to each comment below. If the reviewer should feel that we have not sufficiently responded to his/her comments, we welcome any suggestion that allows us to further improve the manuscript.
Line 76. "Do we need more research?" There is not an answer at the end of the section.
We thank the reviewer for pointing out that our message here may not have been entirely clear. We have rephrased the section and hope it is now more evident. Importantly, as research always generates new questions, we will never reach the point where all questions have been answered rendering further research unnecessary. While we may reach a point, where existing evidence is (even without further research) solid enough to give a certain recommendation (e.g., we can recommend people to not smoke), this is not the case in the DST/ST debate. In such cases, we must not adopt a one-sided perspective, but communicate existing and missing evidence in a balanced manner.
The section now reads:
First, like a perpetuum mobile, research will naturally always generate new questions and thus result in athe need for more research. Critically though, this objection seems to imply be used to argue that in cases where evidence from existing research is not solid, adopting a one-sided perspective was justified. (cf. line 82 et seqq.)
Line 113. "the correlation is positive" is erroneous. It is impossible to obtain an overall negative correlation if all the individuals have a positive one. It is more appropriate "the correlation could be positive".
We are happy to clarify this point. Indeed, a correlation can be negative on the group level but positive when every participant is considered individually. On the group level, every individual contributes just his/her mean speed and the mean number of typos whereas the correlation on the level of the individual is based on the single trials of an individual.
To illustrate this point, it may be helpful to consider the following figure. Here, an individual makes the more mistakes/typos the faster s/he writes. However, on the group level excellent typists are faster and make less mistakes than inexperienced typists.
Line 115. "Even odds ratios do not provide an intuitive understanding" What is intuitive? It is an universal property that not depends on the culture or academic level of the reader or the ability of the lecturer? Hence, theory of relativity is invalidated if it is not intuitively exposed?
We have adapted the section and replaced the term “intuitive” by “a simple quantification”. The section now reads:
Even odds ratios do not provide a simple quantification n intuitive understanding of the risk of an individual, wherefore they are often misinterpreted (Holcomb Jr, Chaiworapongsa, Luke, & Burgdorf, 2001). Additionally, theyand often frequently magnify an existing effect (Holcomb Jr et al., 2001). (cf. lines 118 et seqq.)
Lines 137. "however recently claimed that it was 'biologically large'". That someone says something, it does not prove any kind of crisis in a field. It would be a good idea to gather other works in the same are and recompile their conclusions about the statistical size.
We thank the reviewer for this suggestion. First, it seems important to note that the issue we would like to make the reader aware of at this point is not the size of the statistical effect per se. Rather, it seems important to discuss the fact that a statistical effect size does not necessarily translate to biological or clinical meaningfulness directly and it certainly does not when the scaling of the underlying measure (in this case SJL) is entirely unclear. Furthermore, to the best of our knowledge, the study by Borisenkov and colleagues is the only one directly relating DST/ST to social jetlag, wherefore we cannot provide further estimates of the statistical effect size from other studies. Therefore, we hope that the reviewer is fine with the section remaining as it was.
Line 198. "In addition to this small and biased sample". Why? The authors have analyzed the power of the analysis or the samples used comparing them with the population?
We now make more explicit that the sample was small relative to the size of the European population (<1% as stated in line 219). The sentence now reads:
“In addition to this relative to the size of the European population small and biased sample […]”. (cf. line 222)
The conclusion that the sample was biased first of all arises from the fact that 70% of the participants in the EU-wide survey were German citizens. As pointed out by the Harding Centre for Risk Literacy, the sample was moreover a convenience sample, where representativeness is extremely hard to obtain. Probably, those, who were not interested in the topic, wanted to keep the status quo, or simply did not care about the outcome did not participate. Thus, we think that the conclusion that the sample was biased, is justified.
Line 200. "This survey and its results are highly unrepresentative". Therefore, the survey must be rejected? If it is the case, some numbers must precede this conclusion. Saying something is not the same as proving it.
We thank the reviewer for this remark and fully agree that this does not mean the results should be rejected. We therefore explicitly state at the beginning of the section that “Beyond a necessary scientific discussion, the DST/ST debate should not just ignore the public preference and scientists should not simply try to ‘explain it away’” (cf. line 213). Indeed, an earlier reviewer had requested that we again explicitly state that the survey was unrepresentative. As this, however, follows from what we previously say, we have removed the sentence “Thus, this survey and its results are highly unrepresentative.”.
Line 231. "...the current state of research suggests that the risk of negative effects associated". This statement is written in the Conclusions sections. It is impossible to conclude this from the previous sections. If there are contradictory researches, some of them shown in the text, why give the authority to one faction?
We admit that it is beyond the scope of this manuscript to give a full review of the state of research, other publications have attempted to do this (e.g. Roenneberg et al., 2019a; Roenneberg et al., 2019b; Dijk et al., 2018). However, in the section entitled “Current consensus position” (cf. p. 3), we summarise the conclusions that result from the current state of research. From this perspective, the sentence in the Conclusion section the reviewer is pointing to is a short summary of the section on p. 3. Besides this, we explicitly try to provide a balanced perspective and not favour one interpretation of the findings.
Line 239. "...Therefore, we encourage a discourse, which aims at getting the public on board..." Before the Conclusions, there is not any paragraph discussing about this issue. In the Abstract is written in Line 27 "therefore recommend strategies that could support a discourse aiming". There are not strategies or section talking about them.
We agree that the section title “Conclusions” may be a little misleading as it is not just a summary, but includes new ideas that logically follow from what has previously been said. We have therefore decided to call this section “Concluding remarks” and hope it now better reflects its content.
Beyond this, we expand a little on the strategies and suggest that the think tanks may be “led by independent institutions such as for example the National Academy of Sciences Leopoldina in Germany”. We would like to explicitly state that this is just one example for such an institution and the suggestion does not result from a political strategy or bias. In an earlier version of the manuscript, one reviewer felt we might be politically biased, wherefore we had removed this suggestion.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This opinion paper discusses the input of the scientific community regarding the debate about standard vs. daylight saving time and the form the latter should take to contribute to future action plans for a potential abolishment of the bi-annual clock change (in Europe).
Since autumn 2018 (announcement of the European Commission President that the bi-annual clock change in Europe should be abolished), a series of opinion papers have emerged; some of them with the intention to form a consensus across scientists (e.g. Consensus Summary of the Panel Discussion at the Meeting of the European Sleep Research Society 2018).
The paper highlights a series of general misconceptions (e.g. the danger of a one-sided perspective when clear research evidence is lacking, “statistical” inferences at the individual level and the interpretation of effect sizes) which are of course worth being highlighted within this context.
At first glance, the paper appears as a response of the authors to a publication in Frontiers in Physiology (Roenneberg, Winnebeck & Klerman, 2019). In the latter Roenneberg et al. report a series of “fallacies” more or less directly addressed to the authors of the present paper. A reply is well funded, but the question arises of whether it wouldn’t be more appropriate to publish the opinion as a comment/response in Frontiers.
I suggest avoiding claims such as:
“A German researcher warned that perennial DST would cause “enormous problems” (translation of the authors, German original: “riesige Probleme”) and render us “fatter, more stupid, and grumpier” (translation of the authors, German original: “dicker, dümmer und grantiger”).
The opinion of the authors (i.e. not fanning fear; providing a balanced view) appears strongly, even without such claims.
The authors discuss autumn and spring time effects on myocardial infarction and the possibility that DST/ST effects are balanced out at the group level. I agree that group-level analyses are not suited for inferences at the individual level. However, does the reasoning necessarily illegitimate the possibility that "autumn relief" does not rescue a"spring victim" (as proposed in Roenneberg et al., 2019) and thus that outbalancing effects at the group level do not imply that this will be (or not be) the case at the individual level?
p.5-6: The authors mention that we lack information about the practical meaningfulness of differences in SJL on an interval scale. What do the authors mean with “practical meaningfulness on an interval scale”? While I appreciate the DALY approach and see its general utility in the debate , I do not see how adding it would solve the issue about interpreting effect sizes per se. The authors might clarify this point and eventually briefly refer to the literature relating SJL with "risk factors" which might be a starting point for the discussion here.
The authors mention the results of a survey and emphazise on the fact that most respondents in favour of perennial ST (43%) mentioned general health considerations the main argument (43%), the main argument (42%) for perennial DST were evening leisure time activities. It might be worth reminding here that summer time inherently goes along with longer days.
Modelling data suggest that the effect of a later solar noon under DST is modulated by artificial light consumption (cf. Figure S3 Skeldon, Phillips, & Dijk, 2017): the higher light levels were in the evening, the less strong were the effects of a shift in solar noon (expressed by longitude differences) on SJL. This is an interesting observation which adds to the existing reports. It might be useful for the reader to add the Figure highlighting this to make the point more clear and relevant for the readership.
p.7, line 181: the authors cite Shochat et al., 2019 “individual timing of sleep and activity in a modern environment with electrical lighting largely conforms to clock time and actual light exposure (which is the combination of natural and electrical light exposure) rather than sun time indexed by the timing of SN (solar noon, A/N)”. In the same publication, Shochat (2019) also state that “although our findings show that individually determined daily light exposure is more strongly associated with sleep timing than merely sun time, they do not imply causality…Yet even though light exposure may be driven by (rather than driving) the sleep-wake cycle, which in turn may be driven by social constraints, this light exposure will have an impact on the circadian pacemaker. This information might also be worth mentioning to complete the picture here.
Author Response
This opinion paper discusses the input of the scientific community regarding the debate about standard vs. daylight saving time and the form the latter should take to contribute to future action plans for a potential abolishment of the bi-annual clock change (in Europe).
Since autumn 2018 (announcement of the European Commission President that the bi-annual clock change in Europe should be abolished), a series of opinion papers have emerged; some of them with the intention to form a consensus across scientists (e.g. Consensus Summary of the Panel Discussion at the Meeting of the European Sleep Research Society 2018).
The paper highlights a series of general misconceptions (e.g. the danger of a one-sided perspective when clear research evidence is lacking, “statistical” inferences at the individual level and the interpretation of effect sizes) which are of course worth being highlighted within this context.
We would like to sincerely thank the reviewer for his/her time and efforts and the helpful and appreciative suggestions. We believe they helped us to substantially improve the manuscript. Below, we answer to each suggestion. If the reviewer should feel that we have not sufficiently responded to his/her comments, we welcome any suggestion that allows us to further improve the manuscript.
At first glance, the paper appears as a response of the authors to a publication in Frontiers in Physiology (Roenneberg, Winnebeck & Klerman, 2019). In the latter Roenneberg et al. report a series of “fallacies” more or less directly addressed to the authors of the present paper. A reply is well funded, but the question arises of whether it wouldn’t be more appropriate to publish the opinion as a comment/response in Frontiers.
We thank the reviewer for this suggestion and absolutely agree that the paper would be more appropriate as a comment/response in Frontiers. Indeed, the manuscript had been with Frontiers for several months and underwent a major revision last year. Unfortunately, Frontiers would not publish it unless we adopted the opinion/perspective of one of the reviewers, which clearly contrasted ours. Therefore, we decided to choose a different journal.
I suggest avoiding claims such as:
“A German researcher warned that perennial DST would cause “enormous problems” (translation of the authors, German original: “riesige Probleme”) and render us “fatter, more stupid, and grumpier” (translation of the authors, German original: “dicker, dümmer und grantiger”).
The opinion of the authors (i.e. not fanning fear; providing a balanced view) appears strongly, even without such claims.
We thank the reviewer for this suggestion and have removed the sentence accordingly. We now merely paraphrase the situation to illustrate the two diverging positions. The sentence now reads:
“Some researchers warned against enormous problems this would cause for public health, a perspective that was also well-perceived by the media (Simon, 2018a, 2018b).” (cf. line 38 et seqq.)
The authors discuss autumn and spring time effects on myocardial infarction and the possibility that DST/ST effects are balanced out at the group level. I agree that group-level analyses are not suited for inferences at the individual level. However, does the reasoning necessarily illegitimate the possibility that "autumn relief" does not rescue a "spring victim" (as proposed in Roenneberg et al., 2019) and thus that outbalancing effects at the group level do not imply that this will be (or not be) the case at the individual level?
We agree with the reviewer that, theoretically, it may be true that the spring effects do not balance out the autumn effects on an individual level, which we also mention in the text (cf. line 105). Importantly though, even though theoretically legitimate, we argue that this reasoning would require that myocardial infarction or its prevention can, beyond doubt, be attributed to clock change. As this will hardly ever be possible, such a statement is misleading and, in practice, seems illegitimate.
5-6: The authors mention that we lack information about the practical meaningfulness of differences in SJL on an interval scale. What do the authors mean with “practical meaningfulness on an interval scale”? While I appreciate the DALY approach and see its general utility in the debate, I do not see how adding it would solve the issue about interpreting effect sizes per se. The authors might clarify this point and eventually briefly refer to the literature relating SJL with "risk factors" which might be a starting point for the discussion here.
We thank the reviewer for pointing out that the message of this paragraph was not entirely clear. We have revised it and hope it is now more comprehensible. The section now reads:
“While that may be true, such a claim is unjustified as long as we lack information about the practical meaningfulness of differences in SJL on an interval scale, that is knowledge about how a certain change in of SJL (e.g. of 20, 30, or 40 minutes) relates to a change on an outcome variable (e.g., symptoms of seasonal affective disorder or body mass index). More precisely, sSuch a judgment would require studies that go beyond just correlations and investigate the causal relationships between SJL and outcome variables. and provide knowledge about how certain values of SJL relate to an outcome (e.g., symptoms of seasonal affective disorder).” (cf. line 148 et seqq.)
At the beginning of the section, we have now also added a paragraph on SJL as a risk factor as recommended by the reviewer.
“In the DST vs. ST debate, the only empirical study directly investigating the effects of perennial DST and ST for example found that DST was associated with an increase in “social jetlag” (SJL; Borisenkov et al., 2017) defined as the difference between the mid-sleep point on workdays and free days. Repeatedly, SJL been found to be a risk factor for higher body mass index (BMI) or even obesity, depressive symptoms, and for behaviour that is hazardous to health such as smoking and poor dietary habits (for a review see Roenneberg, Pilz, Zerbini, & Winnebeck, 2019).” (cf. lines 139 et seqq.)
Moreover, would like to clarify that DALYs will not solve the issues related to the interpretation of effect sizes per se. Rather, we suggest they could be an alternative, which allows for an appreciation of the relative practical meaningfulness (i.e., in comparison to well-known risk factors) of the effects of e.g. time zone position. We have tried to make more explicit that DALYs could be an alternative rather than solve problems associated with statistical effect sizes. The sentence now reads:
“We here would also like to suggest an alternative to statistical effect sizes approach to that could help appreciate the practical meaningfulness of certain factors such as time zone position in epidemiological studies.” (cf. line 158 et seqq.)
The authors mention the results of a survey and emphazise on the fact that most respondents in favour of perennial ST (43%) mentioned general health considerations the main argument (43%), the main argument (42%) for perennial DST were evening leisure time activities. It might be worth reminding here that summer time inherently goes along with longer days.
We thank the reviewer for this thoughtful suggestion and have added this to the text. The section now reads “[…] the main argument (42%) for perennial DST were evening leisure time activities, which are facilitated by the prolonged availability of daylight.”(cf. line 128 et seqq.)
Modelling data suggest that the effect of a later solar noon under DST is modulated by artificial light consumption (cf. Figure S3 Skeldon, Phillips, & Dijk, 2017): the higher light levels were in the evening, the less strong were the effects of a shift in solar noon (expressed by longitude differences) on SJL. This is an interesting observation which adds to the existing reports. It might be useful for the reader to add the Figure highlighting this to make the point more clear and relevant for the readership.
This is an excellent suggestion, which we have gladly adopted. We here copy the figure and its caption for convenience. Note that we focus on clock time of the core body temperature minimum and social jetlag as these were the two examples mentioned in the text.
Figure 1. Changes in circadian timing and social jetlag across a time zone for different levels of evening light (range 0-100 lux). (A) Core body temperature minimum (Cmin) indicating the minimum of the circadian wake propensity. Cmin occurs later in the west of a time zone at low levels of evening light as indicated by the smaller slope of the lines. With increasing levels of evening light, the effect of the time zone position disappears as indicated by the steepening slope. (B) Social jetlag is larger in the western parts of a time zone (cf. smaller slope), but decreases with increasing levels of evening light (cf. increasing slope of the lines). Note that a 15° shift to the east (west) within a time zone corresponds to an advance (delay) in solar noon by one hour relative to clock time. Abbreviations: Cmin = core body temperature minimum; hrs = hours. Adapted from Skeldon et al. (2017).
7, line 181: the authors cite Shochat et al., 2019 “individual timing of sleep and activity in a modern environment with electrical lighting largely conforms to clock time and actual light exposure (which is the combination of natural and electrical light exposure) rather than sun time indexed by the timing of SN (solar noon, A/N)”. In the same publication, Shochat (2019) also state that “although our findings show that individually determined daily light exposure is more strongly associated with sleep timing than merely sun time, they do not imply causality…Yet even though light exposure may be driven by (rather than driving) the sleep-wake cycle, which in turn may be driven by social constraints, this light exposure will have an impact on the circadian pacemaker. This information might also be worth mentioning to complete the picture here.
We welcome this suggestion and agree that it should be included to complete the picture. The section now reads:
“Here, the authors conclude that ‘individual timing of sleep and activity in a modern environment with electrical lighting, largely conforms to clock time and actual light exposure (which is the combination of natural and electrical light exposure) rather than sun time indexed by the timing of SN (solar noon, A/N)’. Note that, although the findings do not imply causality and the timing of sleep also influences the timing of light exposure, any light exposure will impact the circadian pacemaker and thereby sleep timing (Skeldon et al., 2017; Wright et al., 2013).” (cf. line 206 et seqq.)
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The authors have satisfactorily answered all my questions. In the answers, it can be observed a sincere attempt to answer the questions raised.
