Those who tan and those who don’t: A natural experiment on colorism (2024)

Tamar Kricheli Katz, Conceptualization, Formal analysis, Investigation, Methodology, Project administration, Supervision, Writing – original draft, Writing – review & editing,^#1,* Tali Regev, Conceptualization, Formal analysis, Investigation, Methodology, Project administration, Writing – original draft, Writing – review & editing,^#2 Shay Lavie, Methodology, Writing – original draft, Writing – review & editing,¹ Haggai Porat, Data curation, Formal analysis, Investigation, Writing – review & editing,³ and Ronen Avraham, Investigation, Methodology, Resources, Writing – review & editing¹

Table 1

lnpit1−pit=skintoneiβ+δXit+φt+εit,

Skintoneit=α+UVitβ+δXit+εit

lnpit1−pit=α+UVitβ+δXit+γi+φt+εit,

Table 2

Table 3

Women and men

In the models presented in Table 3, we explore the differences between the effects of the average UV radiation in the previous three weeks on women and on men (with intermediate skin tones).

We find that the effect of the average UV radiation in the previous three weeks is negative for men but not for women. On average, a one unit change in the average UV radiation in the previous three weeks leads to a 0.913 fold decrease in the odds of being employed for male respondents with intermediate skin tones (and who therefore tend to tan) (p<0.001). One possible explanation for this finding is that it is discrimination against men that drives the results we observe. This possibility is consistent with the results of previous studies that have shown that race inequality in the American labor force is driven by differences between white and black men [19, 26, 27]. Another possible explanation is that women may be more likely than men to use sunscreen. Indeed, some studies suggest that women tend to use more sunscreen on the face than men. However, these studies also show that most adults (women and men) do not regularly use sunscreen on the face [28]. In any case, note that if a large number of respondents (women and men) in our dataset tended to use sunscreen, we would be underestimating the effects of UV radiation on employment in our analyses. As expected, the effects of the average UV radiation in the previous three weeks on the employment of men with the lightest and darkest skin tones are statistically non-significant.

Sun exposure

One possible alternative explanation to our findings is that the people with intermediate skin tones tend to work in occupations in which employment is lower when UV radiation is higher. To rule out this explanation, in S3 Table in the S1 Appendix (Model 1), we include an Occupational Information Network’s (O*NET) index variable that captures the degree to which one’s occupation typically involves exposure to the sun (ranging from 1 to 100, at the 4-digit Standard Occupational Classification). In a separate analysis, we find that people with intermediate skin tones tend to work more in occupations that involve greater sun exposure. Because we found that effects are only significant for men, we present the results only for men.

We find that even after including the sun-exposure-index variable, the effect of the average UV radiation in the previous three weeks remains negative and significant for people with intermediate skin tones and non-significant for all others. We also find that the interaction between the sun exposure index variable and the average UV radiation in the previous three weeks is actually positive. This suggests that men are more likely, not less, to be employed in occupations that involve greater exposure to the sun when the average UV radiation increases. We therefore conclude that it is not the tendency of people with intermediate skin tones to work in occupations that involve exposure to the sun that drives our results.

Place of residence

Another alternative explanation for the results we find is that the average UV radiation in the previous three weeks is correlated with one’s place of residency that in turn affects employment in ways that are not observable in our dataset (and that are relevant only to people with intermediate skin tones). One such possible explanation might be that people with intermediate skin tones move to sunny places in which they tend to be employed less. To rule out this explanation, we report results of a logistic regression model predicting the effects of the average UV radiation in the previous three weeks on one’s employment for the sample of men who have not yet changed their place of residency. By doing so, we keep respondents’ place of residence constant. We find that even when the sample includes only observations in which the place of residence is constant, the effect of the average UV radiation in the previous three weeks on one’s probability of employment is negative and statistically significant although much smaller; On average, a one unit change in the average UV radiation in the previous three weeks leads to a 0.932 fold decrease in the odds of being employed (p<0.001; Model 2, S3 Table in the S1 Appendix). This is a smaller effect compared to the effect of UV radiation in the previous three weeks on employment for all the men in our sample (0.913, p<0.001, Model 1, Table 3).

Occupations with greater turnover

The design of our study enables us to capture discrimination only when people’s employment statuses change. In other words, the more people’s employment statuses change, the easier it would be for us to detect the effects of the average UV radiation in the previous three weeks on the probability of being employed. Therefore, we expect the effect of the UV radiation in the previous three weeks to be more detectible when respondents are employed in occupations with greater turnover. In Model 3 (S3 Table in the S1 Appendix), we therefore include an index variable capturing the duration of employment per occupation. This variable was constructed from the Current Population Survey (CPS) as the average duration of employment in the current position (in years), by occupation (ranging from 0.24 to 26, by year, at the 4-digit Standard Occupational Classification). We find that indeed, the effect observed in our analysis for the average UV radiation in the previous three weeks is greater when the duration of employment within the occupation is shorter. Note also that occupations with greater turnover tend to be low-status occupations. Indeed, in our data effects were stronger for low-status occupations. This may be because of the above-mentioned effect of turnovers but also because with low status occupations the length of time between interview and employment is shorter compared to with high status occupations (where oftentimes the interview takes place weeks before the change in employment status happens). In other words, with high status occupations (compared to low-status occupations) looking darker when first employed is less strongly correlated with being darker when interviewed. Thus, the design of our study makes it harder to detect discrimination on the basis of one’s perceived skin tone in high-status occupations compared to low status occupations.

Robustness checks and additional inquires

To check the robustness of the results we report, we try alternative specifications to test our hypothesis (S4 Table in the S1 Appendix). We use the average UV radiation in the previous four weeks (instead of the previous three weeks) (model 1); we use different categorizations of respondents’ skin tones (models 2–4); and we test for nonlinear effects of the average UV radiation in the previous three weeks (model 5). In all alternative specifications, the results obtained are essentially the same with negative and statistically significant effects of the average UV radiation on the probability of employment for respondents with intermediate skin tones, but not others.

We also estimate the effects of longer lags of UV radiation. We find that other specifications are also negatively correlated with the current employment of people with intermediate skin tones. As predicted, for people with lighter or darker skin tones, effects on employment are statistically non-significant (models are estimated on male non movers whose skin tones were assessed in days in which the average UV radiation in the previous three weeks was relatively high).

In addition, we explore the effects of the season in which respondents’ skin tone was assessed on the results we obtain. Recall that the skin tones of respondents were assessed on different dates in 2008. Whereas the skin tone of many respondents was assessed in the winter, the skin tone of some was assessed in May or in October when UV radiation tends to be greater (none were interviewed during the June-September period). Because some skin tones are affected by UV radiation, we worry that the assessed skin tones of the respondents that were interviewed in May or October are darker than they would have been if assessed in the winter, generating inaccurate classification of the three skin tone categories we use. More specifically, we are concerned that those assessed in May and October as having intermediate skin tones, actually have lighter skin tones than those assessed as having intermediate skin tones in the winter.

In Models 1 and 2 presented in S5 Table in the S1 Appendix, we therefore separate those respondents (with intermediate skin tones) whose skin tone was assessed in the winter from those assessed in May or October. Indeed, we find that the general effects for the average UV radiation in the previous three weeks we observe are driven by the people whose skin tone was assessed in the winter, and not in May or October where exposure to UV radiation might have affected the assessment of skin tone.

We further explore the effects on employment of the average UV radiation in the future (weeks 5–7) together with the average UV radiation in the previous three weeks (on people with intermediate skin tones), as a falsification test (Model 3, S5 Table in the S1 Appendix). We find that it is only the UV radiation in the previous three weeks but not in the following three weeks that affects one’s employment. We also estimate the effects of ten data points of UV radiation (weeks 1–10) in the future on past employment outcomes (together with the average UV radiation in the previous three weeks; on people with intermediate skin tones). All ten future data points estimated had statistically non-significant effects on past employment. Next, we estimate the effects of the average UV radiation in the previous three weeks within race. We find that even within race, greater UV radiation in the previous three weeks is associated with lower probability of employment (Models 4–6, S5 Table in the S1 Appendix).

In model 4, we see that even within whites, the effects of UV radiation on employment are negative and statistically significant. Recall that in our sample (Table 1) 35% of the people with intermediate skin tones were whites. These findings suggest that even within this white population looking darker is associated with a lower probability of being employed. This highlights the negative effects of colorism even within race, and perhaps surprisingly even within whites.

Estimations of time trends and variations by geographical regions are all statistically non-significant.

Note that concerns of reversed causality (i.e, that unemployed individuals spend more time in the sun and thus look darker) are irrelevant for our research design; we do not observe how dark people currently look nor how much time they actually spend in the sun. We only proxy their potential to look darker by using their assessed skin tone in 2008 and the (exogenously given) average UV radiation in the previous three weeks. In other words, being unemployed in the present cannot affect the assessed skin tone in 2008 or the average UV radiation in the previous three weeks.

Moreover, in our data set, the assessed skin tones of unemployed participants are not affected more by the average UV radiation in the previous three weeks than the assessed skin tones of the employed participants. In models predicting the assessed skin tone by the average UV radiation in the previous three weeks (similar to the models presented in S2 Table in the S1 Appendix), an interaction between being unemployed and the average UV radiation in the previous three weeks is statistically non-significant.

Finally, an additional set of possible concerns would be that people who look like they spent time in the sun are perceived to be lazier or less committed to the labor force than people who do not, and that the effects we observe are the results of these perceptions. Note however, that people with intermediate skin tones (unlike those with very light skin tones) look as if their skin tone is originally darker after spending time in the sun (see the results presented in S2 Table in the S1 Appendix). More importantly, there is no reason to assume that only people with intermediate skin tones who spend time in the sun are perceived to be lazy or less committed, but not people with the lightest skin tones who spend time in the sun.

Interestingly, one recent study has shown that when the weather is nicer, workers are more likely to report being sick [29]. This should not bias our results however, because in the NLSY97 dataset sickness or absenteeism are not coded as unemployment.

S1 Appendix

(DOCX)

1. Lang K., & Manove M. (2011). Education and Labor Market Discrimination. American Economic Review, 101, 1467–1496. [Google Scholar]

2. Eckstein Z., & Wolpin K. (1999). Estimating the Effect of Racial Discrimination on First Wage Job Offers. Review of Economics and Statistics, 81, 384–392. [Google Scholar]

3. Neal Derek, and Johnson William, “The Role of Premarket Factors in Black-White Wage Differences,” Journal of Political Economy104 (1996), 869–895. [Google Scholar]

4. Ritter Joe; Taylor Lowell J. / Racial disparity in unemployment. In: Review of Economics and Statistics. 2011; Vol. 93, No. 1 pp. 30–42. [Google Scholar]

5. Bertrand M., & Mullainathan S. (2004). Are Emily and Greg More Employable than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination. American Economic Review, 94, 991–1013. [Google Scholar]

6. Pager D. (2003). The Mark of a Criminal Record. American Journal of Sociology, 108(5), 937–975. [Google Scholar]

7. Pager D., Western B. & Bonikowski B. (2009). Discrimination in a Low-Wage Labor Market: A Field Experiment. American Sociological Review, 74, 777–799. 10.1177/000312240907400505 [PMC free article] [PubMed] [CrossRef] [Google Scholar]

8. Kirschenman J., & Neckerman K. (1991). We'd love to hire them but…: The Meaning of Race to Employers In Jencks C. & Peterson P. (Eds.), The Urban Underclass (pp. 203–232). Washington, D.C: Brookings. [Google Scholar]

9. Holzer H. (1999). What Employers Want: Job Prospects for Less-Educated Workers. New York: Russell Sage Foundation. [Google Scholar]

10. Moss P. & Tilly C. (2001). Stories Employers Tell: Race, Skill, and Hiring in America. New York: Russell Sage. [Google Scholar]

11. Waldinger R. & Lichter M. (2003). How the Other Half Works: Immigration and the Social Organization of Labor. Berkeley, CA: University of California Press. [Google Scholar]

12. Harrison M. S., & Thomas K. M. (2009). The hidden prejudice in selection: A research investigation on skin color bias. Journal of Applied Social Psychology, 39(1), 134168. [Google Scholar]

13. Goldsmith A., Hamilton D., & Darity W. (2006). Shades of Discrimination: Skin Tone and Wages. American Economic Review, 96(2), 242–45. [Google Scholar]

14. Hughes M., & Hertel B. (1990). The Significance of Color Remains: A Study of Life Chances, Mate Selection, and Ethnic Consciousness Among Black Americans. Social Forces, 68(4),1105–20. [Google Scholar]

15. Monk Ellis P.2015. “The Cost of Color: Skin Color, Discrimination, and Health among African Americans.” American Journal of Sociology121(2):1–42. [PubMed] [Google Scholar]

16. Kreisman D. & Rangel M. (2015). On the Blurring of the Color Line: Wages and Employment for Black Males of Different Skin Tones. The Review of Economics and Statistics, 97(1), 1–13. [Google Scholar]

17. Keith Verna M., Herring Cedric. (1991). “Skin Tone and Stratification in the Black Community.” American Journal of Sociology97(3):760–78. [Google Scholar]

19. Hersch J. (2006). Skin tone effects among African Americans: Perceptions and reality. American Economic Review, 96(2), 251–255. [Google Scholar]

20. Neal D. & Johnson W. (1996). The Role of Premarket Factors in Black-White Wage Differences. Journal of Political Economy, 104(5), 869–95. [Google Scholar]

21. Cosmides L., Tooby J., & Kurzban R. (2003). Perceptions of race. Trends in Cognitive Sciences, 7, 173–179. 10.1016/s1364-6613(03)00057-3 [PubMed] [CrossRef] [Google Scholar]

22. Fitzpatrick T. B. (1975). "Soleil et peau" [Sun and skin]. Journal de Médecine Esthétique(in French), (2), 33–34. [Google Scholar]

23. Fitzpatrick T.B. (1988). The validity and practicality of sun-reactive skin types i through vi.Archives of Dermatology,124(6), 869–871. [PubMed] [Google Scholar]

24. Chardon A., Cretois I., & Hourseau C. (1991). Skin Colour Typology and Suntanning Pathways. International Journal of Cosmetic Science, 13, 191–208. 10.1111/j.1467-2494.1991.tb00561.x [PubMed] [CrossRef] [Google Scholar]

25. Fernández-Val Iván & Weidner Martin, (2016). "Individual and time effects in nonlinear panel models with large N, T," Journal of Econometrics, vol. 192(1), 291–312. [Google Scholar]

26. Hadas M., & Semyonov M. (2016). Going Back in Time? Gender Differences in Trends and Sources of the Racial Pay Gap, 1970–2010. American Sociological Review, 81(5),1039–1068. [Google Scholar]

27. Navarrete C. D., McDonald M. M., Molina L. E., & Sidanius J. (2010). Prejudice at the nexus of race and gender: an outgroup male target hypothesis. Journal of personality and social psychology, 98(6), 933 10.1037/a0017931 [PubMed] [CrossRef] [Google Scholar]

28. Holman D. M., Berkowitz Z., Guy G. P. Jr, Hawkins N. A., Saraiya M. (2015), Patterns of sunscreen use on the face and other exposed skin among US adults. Journal of the American Academy of Dermatology, 73(1):83–92. 10.1016/j.jaad.2015.02.1112 [PMC free article] [PubMed] [CrossRef] [Google Scholar]

29. Shi and Skuterud, 2015, Gone Fishing! Reported Sickness Absenteeism and the Weather, Economic Inquiry53(1), 388–405. [Google Scholar]

26 Mar 2020

PONE-D-20-03529

Those Who Tan and Those Who Don’t: A Natural Experiment on Colorism

PLOS ONE

Dear Dr. Kricheli-Katz,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Both the qualified reviewers are labor economists and they like the novelty of your research. I myself also enjoyed reading your paper. But the reviewers also raised some major concerns. Please try to address their concerns as much as you can. In addition, I also have some comments and suggestions.

First, reviewer 1 is a bit confused about the coefficients versus odds ratio in your table. I suggest you at least report both the coefficients (log odds ratio) and odds ratio in Table 2 and then you can state you will report only odds ratio in the rest of tables. By looking at the negative coefficients for column 3 and 4 of table 2, readers will see your results intuitively.

Second, as reviewer 2 pointed out women may be more likely to use sunscreen. Such kind of avoidance behavior may occur among men too. For example, if people are concerned about the risk of skin cancer, they may be more likely to avoid sunshine on days with high degree of UV radiation. If this is true also for people with intermediate tones, the effect of UV on employment is biased toward zero. This may be true if people with intermediate tone understand they may be more likely to be discriminated if their skin get darker.

Reviewer 2 also raised a valid question on the tradeoff between nice weather and leisure or work. Actually there is a study showing that when weather is nice workers are more likely to report sick absence (Shi and Skuterud, 2015, Gone Fishing! Reported Sickness Absenteeism and the Weather, Economic Inquiry 53(1), 388-405.) You may cite this paper. One way to address this concern is to check full-time workers versus part time workers. UV radiation is expected to affect full time employment less but part time employment more if people do value leisure on sun shining days. In this case you may use an ordered logit model, using 0 as unemployment, 1 as part time, 2 as full time.

Third, the within-race results in columns 4-6 of Table A4 is tricky to interpret. On the one hand, they show your estimation results are robust; on the other hand, why would the UV effect hold for white people assuming white people are not discriminated and have light tone. Does Column 4 suggest a “tan” effect—people who got a tan are more likely to be unemployed? Since your models have included individual fixed effects which absorb the racial effect, I tend to suggest you drop these columns unless you have a better explanation for the white sample.

One minor question out of my curiosity is how many days does it take to get a tan and how long does a tan persist? Are there any medical or scientific studies on these questions?

We would appreciate receiving your revised manuscript by May 02 2020 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

• A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

• A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

• An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Shihe Fu, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

1. When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have indicated that data from this study are available upon request. PLOS only allows data to be available upon request if there are legal or ethical restrictions on sharing data publicly. For more information on unacceptable data access restrictions, please see http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions.

In your revised cover letter, please address the following prompts:

a) If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent.

b) If there are no restrictions, please upload the minimal anonymized data set necessary to replicate your study findings as either Supporting Information files or to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories.

We will update your Data Availability statement on your behalf to reflect the information you provide.

3. We note that Figure [1 A] in your submission contain copyrighted images. All PLOS content is published under the Creative Commons Attribution License (CC BY 4.0), which means that the manuscript, images, and Supporting Information files will be freely available online, and any third party is permitted to access, download, copy, distribute, and use these materials in any way, even commercially, with proper attribution. For more information, see our copyright guidelines: http://journals.plos.org/plosone/s/licenses-and-copyright.

We require you to either (1) present written permission from the copyright holder to publish these figures specifically under the CC BY 4.0 license, or (2) remove the figures from your submission:

1. You may seek permission from the original copyright holder of Figure [1 A] to publish the content specifically under the CC BY 4.0 license.

We recommend that you contact the original copyright holder with the Content Permission Form (http://journals.plos.org/plosone/s/file?id=7c09/content-permission-form.pdf) and the following text:

“I request permission for the open-access journal PLOS ONE to publish XXX under the Creative Commons Attribution License (CCAL) CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). Please be aware that this license allows unrestricted use and distribution, even commercially, by third parties. Please reply and provide explicit written permission to publish XXX under a CC BY license and complete the attached form.”

Please upload the completed Content Permission Form or other proof of granted permissions as an "Other" file with your submission.

In the figure caption of the copyrighted figure, please include the following text: “Reprinted from [ref] under a CC BY license, with permission from [name of publisher], original copyright [original copyright year].”

2. If you are unable to obtain permission from the original copyright holder to publish these figures under the CC BY 4.0 license or if the copyright holder’s requirements are incompatible with the CC BY 4.0 license, please either i) remove the figure or ii) supply a replacement figure that complies with the CC BY 4.0 license. Please check copyright information on all replacement figures and update the figure caption with source information. If applicable, please specify in the figure caption text when a figure is similar but not identical to the original image and is therefore for illustrative purposes only.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Referee Report

PONE-D-20-03529

Those Who Tan and Those Who Don’t: A Natural Experiment on Colorism

Summary:

This paper looks for the evidence of colorism by studying the effect of skin tone on individual’s employment. Combining individual-level weekly employment data from NLSY97 with the data of UV radiation collected from weather stations, the paper explores whether the exposure to UV radiation during the preceding three weeks affects an individual’s employment. In particular, the main results are obtained for three skin tone categories separately (light, intermediate, dark). Based on the logistic regression model controls for individual fixed effects, the effect is only found statistically significant for the intermediate skin tone category, who are most likely to get darker under UV radiation.

General Comments:

This paper novelly exploits the weekly variations in UV radiation, which potentially shifts an individual’s skin tone, and individual’s employment to empirically test colorism. Moreover, instead of basing on cross-sectional variations, this paper increases credibility by using within-individual variations and examines the heterogeneous effects by different skin tone categories. This paper is also rich in details, taking care of several relevant concerns. Nevertheless, there are several crucial mistakes or confusions needed to be clarified before drawing convincing conclusions.

a. The most concerning and confusing results are the main ones reported in Table 2. The paper claims that UV radiation reduces the likelihood of being employed. However, all coefficients of UV are positive. If the effect is indeed negative, there are three possibilities to explain the results here: (1) dependent variable is an indicator of unemployment instead of employment. (2) Greater value in the UV variable associates with lower UV radiation level. (3) The estimates are negative but the authors report them with typos. The first possibility is against the results in Table 4. For Table 4, the authors claim in page 18 that people “tend to work more in occupations that involve greater sun exposure”, which suggests indeed the dependent variable equals 1 for employment instead of unemployment. The second possibility is against the description of the UV variable in page 6. The third possibility is unlikely as these “typos” exist in every columns in Table 2, Table 3, Table4 as well as Table A1. If these were indeed typos, mistakes are grave and they undermine the rigor of this work.

b. Please report standard errors for all coefficients instead of only indicating the significance level, especially for Table 2 and Table3. This is quite important as the main conclusion that UV has negative effect only for the intermediate tones relies heavily on the statistical significance. Examining the results in Table 2, actually coefficients have close magnitudes for all three skin tone categories.

c. Still related to the similar magnitudes of coefficients in all three skin tone categories, the difference in sample size across categories is a concern: intermediate tones have significantly larger sample size. While the authors claim that the same results are found for a random smaller sample for the intermediate tones, please report the numerical results (the coefficients, the sample size, and the standard errors).

d. Coefficients of almost all variables, regardless of the significance, have the magnitude between 0.9 and 1.0 in the regressions of employment. Consider that different variables have quite different units and scales, such as the variables UV, UV^2 and UV^3 in Table A3, or the variable Sun Exposure Index in Table 4, it is unsure whether this is just a coincidence or not (Are these variables normalized? ). The paper suggests that the main results are based on the individual fixed-effects logit model. Is this the conditional logit model, or just a standard logit regression with individual dummies? The latter one may suffer from the issue of incidental parameters and the estimates are inconsistent. For easier and clearer interpretations, can the authors report results based on the linear probability model with individual fixed effects (at least for Table 2)?

Other Minor Comments:

a. How many individuals are included in the sample? Page 6 suggests the number is 4020, while the descriptive statistics in Table 1 by summing up the row Unique Individuals suggests less.

b. Related to the interpretation of the fixed-effects logit regressions, why the effect of Sun Exposure Index of each occupation can be identified in Table 4? Is this Index provided by O*NET longitudinal? To the best of my knowledge, O*NET provides cross-sectional information for each occupation. How is the effect of this time-invariant variable identified in the fixed-effects model?

c. How does this paper determine the occupation, and occupation-related variables such as the Sun Exposure Index and the Employment Duration, for those individuals unemployed? These variables should be unavailable if an individual is unemployed. Particularly, if the individual has changed occupations across weeks, is there weekly occupation information?

d. The row Week, State, Industry and Occupation Dummies in Table A2 may miss the Y for models (2) and (3).

e. (Pseudo) R-squares should be reported in all regressions.

f. Temperature may be a factor correlates with both the UV and individual’s employment.

Reviewer #2: Referee Report for Manuscript No. PONE-D-20-03529 “Those Who Tan and Those Who Don’t: A Natural Experiment on Colorism”

This manuscript provides evidence on the effect of one’s skin tone on his or her probability of being employed, use exposure to UV radiation as a natural experiment. The authors find male and those with intermediate skin tones are less likely to be employed when the UV radiation in the previous three weeks in the area in which they reside is greater. This is an interesting paper, and below are my comments.

Main comments:

1. The main question, as mentioned in the abstract, is “are darker-skinned workers discriminated against in the labor market?”. However, I don’t think the results found in the paper can fully separate discrimination from alternative hypotheses. For example, if greater UV radiation makes people want to enjoy life more and reduce their job search effort or work effort in the labor market, we would find the same negative correlation between UV radiation and employment probability. Although the authors did mention this possibility in the manuscript (last paragraph in page 23), they argue that only people with intermediate skin tones finds this significant negative effect, but not other people, can rule out the work effort hypothesis. I do not completely agree. It is still possible that intermediate skin tones are related to specific personal characteristics, and those are the people who will reduce their work effort when UV radiation is greater. I would not emphasize too much on a causal discrimination story, and would just tell an interesting correlation between UV radiation and employment, and acknowledge it could due to discrimination or some other reasons.

2. The authors find that UV radiation only has effect on men’s employment but not women’s, and interpret it as “this suggests that it is discrimination against men that drives the results we observe” (page 16, last sentence). Again, one can think of alternative stories other than discrimination. Perhaps women are more likely to use sunscreen and less likely to get darker. Or perhaps women drop out of labor force (e.g., stay home and take care of children) rather than become unemployed. In the latter case, maybe the authors can provide analysis on labor force participation, in addition to employment.

3. Does NLSY surveys every year and the weekly employment status is based on recall? Recalling employment status every week for the past year may not be accurate and is subject to large measurement error. The authors should discuss that.

4. During Christmas season or winter break, many people travel to other city or state and the UV radiation of where they usually reside (I assume the place of reside only reports once each survey year in NLSY) would have no effect on their skin tone. Since the data provide interview dates, perhaps the authors can exclude holiday season where people travel the most, as a robustness check. People can be out of town in other weeks too, but it maybe difficult to detect that.

Minor comments:

5. The authors conduct several solid regression analysis. I would suggest the authors to write out the regression equation explicitly before showing the results.

6. In addition to group the data into three skin tone categories (light, intermediate, dark), I would like to see results in finer categories and not to group them together (perhaps in the appendix), given the data have a large sample size.

7. Table A2 is not mentioned in the text, perhaps it should be referred to in the last paragraph of page 10? What’s the difference between column (1) and (2) in Table A2? Perhaps there is a missing “Y” in the column (2) for dummies, so does column (3)?

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at gro.solp@serugif. Please note that Supporting Information files do not need this step.

5 May 2020

Dear Shihe Fu,

Thank you very much for the valuable comments and suggestions. Following them, we revised the manuscript. Please see the updated manuscript (changes in ‘track changes’) and our responses to the comments below (in Italics).

Please note that that there are legal restrictions on sharing the NLSY97 de-identified data set that contains detailed information on the geographic residence of each NLSY97 respondent. The Geocode CD is released only to those who satisfactorily complete the Bureau of Labor Statistics geocode agreement procedure.

The geocode application document is available online at www.bls.gov/nls/geocodeapp.htm

Because we link the UV radiation data to the geographic residence of respondents, we cannot share the data (but we can share the codes if needed).

Sincerely,

Tamar Kricheli Kat, Tali Regev, Shai Lavie, Haggai Porat and Ronen Avraham.

Dear Dr. Kricheli-Katz,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Both the qualified reviewers are labor economists and they like the novelty of your research. I myself also enjoyed reading your paper. But the reviewers also raised some major concerns. Please try to address their concerns as much as you can. In addition, I also have some comments and suggestions.

First, reviewer 1 is a bit confused about the coefficients versus odds ratio in your table. I suggest you at least report both the coefficients (log odds ratio) and odds ratio in Table 2 and then you can state you will report only odds ratio in the rest of tables. By looking at the negative coefficients for column 3 and 4 of table 2, readers will see your results intuitively.

Following this recommendation, we now report both the odds ratio and the log odds ratio in Tables 2 and 3 and then we state that we report only odds ratio in the appendix tables.

Second, as reviewer 2 pointed out women may be more likely to use sunscreen. Such kind of avoidance behavior may occur among men too. For example, if people are concerned about the risk of skin cancer, they may be more likely to avoid sunshine on days with high degree of UV radiation. If this is true also for people with intermediate tones, the effect of UV on employment is biased toward zero. This may be true if people with intermediate tone understand they may be more likely to be discriminated if their skin get darker.

Thank you for your comment. We now include this alternative explanation in the text. We note however that studies have shown that most adults do not use sunscreen on the face (page 18).

Reviewer 2 also raised a valid question on the tradeoff between nice weather and leisure or work. Actually there is a study showing that when weather is nice workers are more likely to report sick absence (Shi and Skuterud, 2015, Gone Fishing! Reported Sickness Absenteeism and the Weather, Economic Inquiry 53(1), 388-405.) You may cite this paper. One way to address this concern is to check full-time workers versus part time workers. UV radiation is expected to affect full time employment less but part time employment more if people do value leisure on sun shining days. In this case you may use an ordered logit model, using 0 as unemployment, 1 as part time, 2 as full time.

Thank you for the reference. We include it in the paper and discuss it in the text (page 24). Note however that we were not clear enough in explaining that in our dataset sickness or absenteeism are not coded as unemployment. This would mean that people who are absent are treated as employed in our analyses. For this reason, part time (that indeed involves more leisure) is not expected to affect the tendency to be unemployed in nice weather. We hope that the text is clearer now.

Third, the within-race results in columns 4-6 of Table A4 is tricky to interpret. On the one hand, they show your estimation results are robust; on the other hand, why would the UV effect hold for white people assuming white people are not discriminated and have light tone. Does Column 4 suggest a “tan” effect—people who got a tan are more likely to be unemployed? Since your models have included individual fixed effects which absorb the racial effect, I tend to suggest you drop these columns unless you have a better explanation for the white sample.

Following this comment, we added an explanation to the model (4) in the text (page 23). Please Let us know if you still think we should remove the model (and we will). (Note that this is now Table A5, since we moved Table 3 to be A3 in the Appendix)

One minor question out of my curiosity is how many days does it take to get a tan and how long does a tan persist? Are there any medical or scientific studies on these questions?

There are no clear answers to this question. The design of our study builds on these two papers:

Fitzpatrick, T. B. (1975). "Soleil et peau" [Sun and skin]. Journal de Médecine Esthétique (in French), (2), 33–34;

Fitzpatrick, T.B. (1988). The validity and practicality of sun-reactive skin types i through vi. Archives of Dermatology, 124(6), 869–871.

Because affects are unclear we also try the UV radiation in the previous 4 weeks for robustness (p. 21 and Table A4).

Responses to the reviewers

Reviewer #1: Referee Report

PONE-D-20-03529

Those Who Tan and Those Who Don’t: A Natural Experiment on Colorism

Summary:

This paper looks for the evidence of colorism by studying the effect of skin tone on individual’s employment. Combining individual-level weekly employment data from NLSY97 with the data of UV radiation collected from weather stations, the paper explores whether the exposure to UV radiation during the preceding three weeks affects an individual’s employment. In particular, the main results are obtained for three skin tone categories separately (light, intermediate, dark). Based on the logistic regression model controls for individual fixed effects, the effect is only found statistically significant for the intermediate skin tone category, who are most likely to get darker under UV radiation.

General Comments:

This paper novelly exploits the weekly variations in UV radiation, which potentially shifts an individual’s skin tone, and individual’s employment to empirically test colorism. Moreover, instead of basing on cross-sectional variations, this paper increases credibility by using within-individual variations and examines the heterogeneous effects by different skin tone categories. This paper is also rich in details, taking care of several relevant concerns. Nevertheless, there are several crucial mistakes or confusions needed to be clarified before drawing convincing conclusions.

a. The most concerning and confusing results are the main ones reported in Table 2. The paper claims that UV radiation reduces the likelihood of being employed. However, all coefficients of UV are positive. If the effect is indeed negative, there are three possibilities to explain the results here: (1) dependent variable is an indicator of unemployment instead of employment. (2) Greater value in the UV variable associates with lower UV radiation level. (3) The estimates are negative but the authors report them with typos. The first possibility is against the results in Table 4. For Table 4, the authors claim in page 18 that people “tend to work more in occupations that involve greater sun exposure”, which suggests indeed the dependent variable equals 1 for employment instead of unemployment. The second possibility is against the description of the UV variable in page 6. The third possibility is unlikely as these “typos” exist in every columns in Table 2, Table 3, Table4 as well as Table A1. If these were indeed typos, mistakes are grave and they undermine the rigor of this work.

Thank you for this comment. Following your comment, we clarified the tables We now explain when odds ratios and when log odds ratios are used (the coefficients are positive in table 2 only when odds ratios are reported. Because they are smaller than one these are all negative effects). We hope that now the results are reported in a clearer way.

b. Please report standard errors for all coefficients instead of only indicating the significance level, especially for Table 2 and Table3. This is quite important as the main conclusion that UV has negative effect only for the intermediate tones relies heavily on the statistical significance. Examining the results in Table 2, actually coefficients have close magnitudes for all three skin tone categories.

Following your comment, we added the standard errors to tables 2 and 3.

c. Still related to the similar magnitudes of coefficients in all three skin tone categories, the difference in sample size across categories is a concern: intermediate tones have significantly larger sample size. While the authors claim that the same results are found for a random smaller sample for the intermediate tones, please report the numerical results (the coefficients, the sample size, and the standard errors).

We were not sure how to report the results for the small random samples (because they vary from one sample to another and the differences between them are not informative. We do report the sample sizes of the small sizes in the paper: “we randomly selected a group of 583 (301231 observations)”.

d. Coefficients of almost all variables, regardless of the significance, have the magnitude between 0.9 and 1.0 in the regressions of employment. Consider that different variables have quite different units and scales, such as the variables UV, UV^2 and UV^3 in Table A3, or the variable Sun Exposure Index in Table 4, it is unsure whether this is just a coincidence or not (Are these variables normalized? ). The paper suggests that the main results are based on the individual fixed-effects logit model. Is this the conditional logit model, or just a standard logit regression with individual dummies? The latter one may suffer from the issue of incidental parameters and the estimates are inconsistent. For easier and clearer interpretations, can the authors report results based on the linear probability model with individual fixed effects (at least for Table 2)?

See our response to (a). Because we are reporting the odds ratios, the effects are all somewhere between 0.9 and 1 (but see the magnitudes and signs for the log odds ratios coefficients or the marginal effects).

Other Minor Comments:

a. How many individuals are included in the sample? Page 6 suggests the number is 4020, while the descriptive statistics in Table 1 by summing up the row Unique Individuals suggests less.

Thank you for this comment. We do have 4020 unique respondents but there are some missing values for two variables in our data (turnover, exposure to the sun in the occupation). Following the comment, we updated table 1. We hope it better reflects the data now.

b. Related to the interpretation of the fixed-effects logit regressions, why the effect of Sun Exposure Index of each occupation can be identified in Table 4? Is this Index provided by O*NET longitudinal? To the best of my knowledge, O*NET provides cross-sectional information for each occupation. How is the effect of this time-invariant variable identified in the fixed-effects model?

Indeed, the index provided by O*NET is not longitudinal. The effects are generated in the fixed effects models because some participants switched their occupations.

c. How does this paper determine the occupation, and occupation-related variables such as the Sun Exposure Index and the Employment Duration, for those individuals unemployed? These variables should be unavailable if an individual is unemployed. Particularly, if the individual has changed occupations across weeks, is there weekly occupation information?

“The NLSY97 asks respondents age 14 or older to report their occupation for each employer. The question "what kind of work did you do" elicits information on the occupation when the job started. The occupational classification at the job's end date (or at the survey date for on-going jobs) is solicited for all employee jobs lasting more than 13 weeks. Survey staff then code the respondent's occupation at each job.”

This means that unemployed individuals are treated as having their previous occupations.

d. The row Week, State, Industry and Occupation Dummies in Table A2 may miss the Y for models (2) and (3).

Indeed. Thank you. We updated Table A2.

e. (Pseudo) R-squares should be reported in all regressions.

Following your comment, we added the (Pseudo) R-squares to all tables reporting logistic regression results, and (adjusted) R-squares to the table reporting the OLS results.

f. Temperature may be a factor correlates with both the UV and individual’s employment.

Yes. This is probably true. Yet, there is no reason to assume that temperature affects only the employment of individuals with intermediate skin tones (only men actually). Also: the effects of temperature on employment are probably small (if any) when the week (and year) are controlled for.

Reviewer #2: Referee Report for Manuscript No. PONE-D-20-03529 “Those Who Tan and Those Who Don’t: A Natural Experiment on Colorism”

This manuscript provides evidence on the effect of one’s skin tone on his or her probability of being employed, use exposure to UV radiation as a natural experiment. The authors find male and those with intermediate skin tones are less likely to be employed when the UV radiation in the previous three weeks in the area in which they reside is greater. This is an interesting paper, and below are my comments.

Main comments:

1. The main question, as mentioned in the abstract, is “are darker-skinned workers discriminated against in the labor market?”. However, I don’t think the results found in the paper can fully separate discrimination from alternative hypotheses. For example, if greater UV radiation makes people want to enjoy life more and reduce their job search effort or work effort in the labor market, we would find the same negative correlation between UV radiation and employment probability. Although the authors did mention this possibility in the manuscript (last paragraph in page 23), they argue that only people with intermediate skin tones finds this significant negative effect, but not other people, can rule out the work effort hypothesis. I do not completely agree. It is still possible that intermediate skin tones are related to specific personal characteristics, and those are the people who will reduce their work effort when UV radiation is greater. I would not emphasize too much on a causal discrimination story, and would just tell an interesting correlation between UV radiation and employment, and acknowledge it could due to discrimination or some other reasons.

Indeed, we cannot refute the claim about people’s tendency to search less when UV radiation is greater. However, we do not think this should be of a great concern thanks to the fact that the effects are significant only for people with intermediate skin tones. We do think this supports our discrimination hypothesis and so does the fact that effects are significant only for men. Finally, note that the effects of UV radiation on employment are found when the week and the year are held constant.

2. The authors find that UV radiation only has effect on men’s employment but not women’s, and interpret it as “this suggests that it is discrimination against men that drives the results we observe” (page 16, last sentence). Again, one can think of alternative stories other than discrimination. Perhaps women are more likely to use sunscreen and less likely to get darker. Or perhaps women drop out of labor force (e.g., stay home and take care of children) rather than become unemployed. In the latter case, maybe the authors can provide analysis on labor force participation, in addition to employment.

Thank you for this comment. Following your suggestion, we include this additional alternative explanation to the manuscript. We note however that studies have shown that most adults do not use sunscreen on the face (page 18).

3. Does NLSY surveys every year and the weekly employment status is based on recall? Recalling employment status every week for the past year may not be accurate and is subject to large measurement error. The authors should discuss that.

Following the comment, we discuss this in the ‘data and methods’ section of the paper.

4. During Christmas season or winter break, many people travel to other city or state and the UV radiation of where they usually reside (I assume the place of reside only reports once each survey year in NLSY) would have no effect on their skin tone. Since the data provide interview dates, perhaps the authors can exclude holiday season where people travel the most, as a robustness check. People can be out of town in other weeks too, but it maybe difficult to detect that.

We tried running 4 different models (one for each season) testing the effects of UV radiation on employment: The UV radiation in the winter has no effect on the employment of participants (but the UV radiation in the fall, summer and spring does). this may be because: (a) winter break (as suggested). (b) the smaller sample size. (c) people don’t go out when it’s winter (4) people don’t tan (don’t look darker) when UV radiation is very low (like in the winter). We are not sure how to disentangle these 4 different mechanisms.

Minor comments:

5. The authors conduct several solid regression analysis. I would suggest the authors to write out the regression equation explicitly before showing the results.

We added the equations to the paper.

6. In addition to group the data into three skin tone categories (light, intermediate, dark), I would like to see results in finer categories and not to group them together (perhaps in the appendix), given the data have a large sample size.

The dataset is indeed large but has a relatively small number of unique individuals. Using the 1-10 skin tone categories generates very small sample sizes. Instead, we try to group them to different categories, for robustness tests (see the appendix, Table A4).

7. Table A2 is not mentioned in the text, perhaps it should be referred to in the last paragraph of page 10? What’s the difference between column (1) and (2) in Table A2? Perhaps there is a missing “Y” in the column (2) for dummies, so does column (3)?

Yes, following the comment we updated the manuscript and table.

16 Jun 2020

Those who tan and those who don’t: Anatural experiment on Colorism

PONE-D-20-03529R1

Dear Dr. Kricheli-Katz,

Both reviewers are happy with your revision. Reviewer 1 has a careful comment but I think your revision has already addressed it. Therefore, we’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at gro.solp@gnillibrohtua.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact gro.solp@sserpeno.

Kind regards,

Shihe Fu, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Referee Report

PONE-D-20-03529R1

Those Who Tan and Those Who Don’t: A Natural Experiment on Colorism

First of all, I would like to apologize for the misunderstanding on results in Table 2 reported by the previous version of this paper. The authors report the odds ratio rather than the log odds ratio, which explains the negative effects of UV on employment with coefficients greater than 0 but smaller than 1.

The only main concern left is about the larger sample size of the intermediate skin tones. This relates to the important finding of the paper that the negative employment effect is only found statistically insignificant for people with intermediate skin tones, who are mostly prone to tan. The authors may want to dismiss the concern that the statistical significance is due to a larger sample. Indeed, authors mentioned that they have done the work and found similar results based on a randomly-picked subsample of the intermediate tones.

My previous report suggests the authors reporting the results based on these random subsamples (General Comments, bullet c). Perhaps I was not clear enough and the authors do not provide the results in the current version. Actually, a table similar to the Table 2 suffices. The only thing to be done is to show the results based on a random sample of intermediate skin tones with balanced sample size, such as the group of 583 (301231 observations) that the authors have done with. It is good that the authors have tried for different random subsamples of intermediate skin tones. I am not asking to report the results based on all these random subsamples. As they are generated randomly, results based on one of these subsample is enough to show the robustness of the results in Table 2.

Of course, the new Table 2 reports results based on the log odds ratio as well as the standard errors. From these new statistics, we can conveniently calculate the z-values and assess how standard errors matter for the significance. A sample size similar to the lightest tones or darkest tones is likely to produce the standard error similar to these two groups (i.e. 0.11 or 0.12). Under these standard errors, the estimated effects for intermediate tones (i.e. -0.027 and -0.041) will still be significant. Therefore my previous concern about the sample size can be partially addressed.

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

23 Jun 2020

PONE-D-20-03529R1

Those who tan and those who don’t: Anatural experiment on Colorism

Dear Dr. Kricheli Katz:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact gro.solp@sserpeno.

If we can help with anything else, please email us at gro.solp@enosolp.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Shihe Fu

Academic Editor

PLOS ONE