The competing effects of racial discrimination and racial identity on the predicted number of days incarcerated in the US: A national profile of Black, Latino/Latina, and American Indian/Alaska Native populations.

OBJECTIVE: Racial discrimination and racial identity may compete to influence incarceration risk. We estimated the predicted days incarcerated in a national US sample of Black, Latino/Latina, and American Indian/Alaska Native (AI/AN) individuals.
METHODS: We used the 2012-2013 National Epidemiologic Survey on Alcohol and Related Conditions-III (n = 14,728) to identify individual incarceration history. We used zero-inflated Poisson regression to predict the number of days incarcerated across racial discrimination and racial identity scores.
RESULTS: Racial discrimination and identity varied between races/ethnicities, such that racial discrimination exposure was highest among Hispanic individuals, while racial identity was highest among Black individuals. Racial discrimination was positively associated with days incarcerated among Black individuals (β = 0.070, p<0.0001) and AI/AN individuals (β = 0.174, p<0.000). Racial identity was negatively associated with days incarcerated among Black individuals (β = -0.147, p<0.0001). The predicted number of days incarcerated was highest among Black individuals (130 days) with high discrimination scores.
CONCLUSION: Racial discrimination and racial identity were associated with days incarcerated, and the association varied by racial/ethnic sub-group. Informed by these findings, we suggest that intervention strategies targeting incarceration prevention should be tailored to the unique experiences of racial/ethnic minoritized individuals at the greatest risk. Policies aimed at reversing mass incarceration should consider how carceral systems fit within the wider contexts of historical racism, discrimination, and structural determinants of health.

Introduction

Racial/ethnic minoritized people have borne the burden of the mass incarceration phenomenon, with Black populations reaching imprisonment rates nearly 10 times higher than White populations in several states [1]. While the U.S. incarcerated population has slowly declined since 2011 [2], Black, Latino/Latina, and American Indians/Alaska Native (AIAN) populations remain overrepresented in prisons and jails [3].

Several factors are associated with incarceration risk, including poor health [4], substance use and mental health diagnoses [5], poverty [6], and having an incarcerated parent [7]. One factor in particular–racial discrimination–may also be a risk factor for incarceration. The effect of discrimination on involvement in the criminal justice system is often framed in terms of systemic biases operating on higher-order ecological levels [8], but evidence linking interpersonal discrimination with criminal justice outcomes is scant. Importantly, personal experiences with racial discrimination are associated with several social and health sequelae that are closely related to incarceration [9]. Some research has addressed the links between race/ethnicity and discrimination based on attitudes and perceptions toward felony offenders [10,11], but these findings are limited to post-incarceration experiences of other non-racial types of discrimination.

Black populations in general, and Black men in particular, report high levels of lifetime discrimination [12,13]. Black, Latino/Latina, and AIAN populations report more exposure to discriminatory behavior in health care [14] and educational settings [15,16] than White populations. Furthermore, race-related stress and discrimination have broad deleterious effects on multiple biopsychosocial levels [17].

Research addressing racial identity is notably less common than that of discrimination. Broadly, having a strong racial identity–the significance and meaning an individual places on their race–has been shown to mitigate several negative health effects associated with racial discrimination [18]. For example, strong racial identity has been found to buffer the negative effect of discrimination on serious criminal offending [19], but this report was limited to a sample of Black individuals and did not compare offending between racial/ethnic minority groups.

Racial identity may have a protective effect on incarceration risk. Furthermore, racial identity may diminish part of the deleterious effect of racial discrimination on incarceration outcomes. However, little evidence exists regarding the moderating effects of racial identity across levels of racial discrimination in the context of incarceration. Given this knowledge gap, we sought to better understand differential associations between racial discrimination and incarceration, whether these associations are conditional on levels of racial identity, as well as the extent to which these conditions vary between Black, Latino/Latina, and AIAN groups. Specifically, in this paper we 1) describe a nationally representative epidemiological profile of racial discrimination, racial identity, and incarceration history, and 2) estimate the predicted number of days incarcerated across levels of racial discrimination within low, mid, and high levels of racial identity. We also explore differences in the moderating effect of racial identity by race/ethnicity.

Materials and methods

Data source and sample

We used data from the National Epidemiologic Survey of Alcohol and Related Conditions–III (NESARC-III) (April 2012 –June 2013) to assess group differences in incarceration, discrimination, and identity. Our final analytic sample included Black, Latino/Latina, and American Indian/Alaska Native survey respondents (N = 14,728). Given the framing of our study around continued disparities and oppression among these three groups, we did not include White (n = 19,194) or Asian/Native Hawaiian/other Pacific Islander (n = 1,801) respondents in our sample. All racial/ethnic categories were available as a single pre-constructed variable in the original dataset. The development and sampling methods of NESARC-III have been described extensively elsewhere [20]. In short, NESARC-III provides individual-level survey data on topics ranging from substance use and mental health disorders, health services utilization, and many unique social and cultural characteristics. NESARC-III uses a complex sampling design and provides sample weights and strata for use in analyses. Broadly, the survey weights and strata are applied in order to account for the oversampling of housing units in pre-defined high-minority geographic areas. NESARC-III was administered to households within a total of 7,200 segments, and segments were clustered within 150 primary sampling units. Access to NESARC-III was granted to the study team by the National Institute for Alcohol Abuse and Alcoholism. We restricted our analytic sample to Black, Latino/Latina, and AIAN individuals who had complete data for all study variables (n = 14,728).

Variables

Our dependent variable of interest was the total number of days spent incarcerated throughout the lifetime. We defined total number of days incarcerated as the sum of responses to two questions, including: 1) Total duration (days) in jail or juvenile detention center before age 18, and; 2) Total duration (days) in jail or correctional facility since age 18.

We considered racial/ethnic discrimination and racial/ethnic identity as our focal independent variables. NESARC-III discrimination questions were derived from the Experiences with Discrimination scales developed by Krieger and colleagues [21,22] and have been described elsewhere [23]. Participants were asked six questions about lifetime experiences with racial/ethnic discrimination. Discrimination response options were on a 5-point scale and included: 0 = never; 1 = almost never; 2 = sometimes; 3 = fairly often, and; 4 = very often. Discrimination questions addressed how often the respondent experienced discrimination because of their race/ethnicity: 1) in their ability to obtain health care; 2) in how they were treated when they got care; 3) in public; 4) in any other situation; 5) by being called a racist name, and; 6) by being made fun of, picked on, or threatened. We created a discrimination score that represents the frequency, or magnitude, of exposure to discriminatory behavior. We calculated the sum of the six discrimination questions to achieve a score ranging from 0–24.

The NESARC-III racial/ethnic identity scale has been described elsewhere [23]. Broadly, the NESARC-III scale was adopted from previous tools that broadly assessed an individual’s self-concept that derives from their knowledge of, or membership in, a social group [24-26]. Respondents were asked eight questions about their racial/ethnic identity. Identity response options were on a 6-point scale and included: 1 = strongly disagree; 2 = disagree; 3 = somewhat disagree; 4 = somewhat agree; 5 = agree, and; 6 = strongly agree. Identity questions capture the extent to which the respondent: 1) has a strong sense of self as a member of their racial/ethnic group; 2) identifies with other members of their racial/ethnic group; 3) considers most of their close friends to be from their own racial/ethnic group; 4) believes racial/ethnic heritage is important; 5) is more comfortable in social situations where other members of their racial/ethnic group are present; 6) is proud of their racial/ethnic heritage; 7) believes their racial/ethnic background plays a big part in interaction with others, and; 8) believes their values and behaviors are shared by people of their racial/ethnic background. We created an identity scale score by calculating the sum of all eight identity questions to achieve a score ranging from 0–48.

We also considered several relevant covariates in our analyses, based on a priori knowledge of possible confounding in the relationships between incarceration, race/ethnicity, discrimination, and identity. Categorical covariates included sex (male or female); educational attainment (less than high school, high school completion, some college or a two-year degree, or college graduate with a bachelor’s degree), and; lifetime drug or alcohol use disorder (yes or no). We condensed several survey response options to create our new 4-level educational attainment variable. Less than high school included those with no formal schooling or those that completed any grade up through 11. High school completion included those who completed grade 12 or received a graduate equivalency degree (GED). Those with some college or a two-year degree were defined as those who had attended a four-year college but did not receive a bachelor’s degree, those with an associate’s degree, or those with another two-year technical degree. College graduates were defined as those who received a bachelor’s degree, attended some graduate or professional studies (completed bachelor’s degree but not a graduate degree), or completed a master’s degree or equivalent or another higher graduate degree. We also included a rate dependent age variable, defined as the number of days incarcerated divided by the number of years at risk of being incarcerated.

Analysis

We used survey procedures in SAS (v9.4) where appropriate and incorporated sample weights and strata to account for the parent study’s complex sampling design. First, we generated within-group unweighted sample sizes, weighted sample sizes, and weighted percentages for each study variable in the total sample and disaggregated by racial/ethnic group. We also calculated the weighted means for discrimination and identity scores for the total sample and within racial/ethnic groups.

We also estimated the association between discrimination and identity scores and the number of days incarcerated. We grouped discrimination scores (0–24) and identity scores (0–48) into three groups of low, middle, and high based on tertile distributions within each racial/ethnic group. Categorizing into tertiles refers to placing the lowest 33.3% of scores into the low group, the middle 33.3% of scores into the middle group, and the highest 33.3% of scores into the high group. Tertiles were created for identity (Black, low = 0–31, mid = 32–36, high = 37–48; Hispanic, low = 0–32, mid = 33–38, high = 39–48; AI/AN, low = 0–27, mid = 28–33, high = 34–48) and discrimination (Black, low = 0, mid = 0.5–1.5, high = 2–24; Hispanic, low = 0, mid = 0.5–1.5, high = 2–24; AI/AN, low = 0, mid = 0.5–1, high = 1.5–24). Because the majority of our NESARC-III sample were incarcerated for zero days (86%), we used linear regression with days incarcerated defined as following a zero-inflated Poisson (ZIP) distribution [27]. ZIP models solve two regression equations, including 1) a zero logit model that predicts the log odds of being in the zero (no incarceration) group, and 2) a linear model that estimates the predicted days incarcerated. We ran separate models for each race/ethnicity. ZIP models were adjusted for age group, sex, education, and lifetime drug/alcohol use disorder. Finally, to help visualize group differences, we calculated the mean predicted days incarcerated by racial discrimination tertiles and by racial identity tertiles. We plotted mean predicted days at low, mid, and high tertiles for discrimination and identity, and disaggregated our results by race/ethnicity.

Results

Fourteen percent of the sample had ever been incarcerated (Table 1), and the mean days incarcerated for the full sample was 52. AIAN individuals demonstrated the highest proportion of ever been incarcerated (26%), while Black individuals demonstrated the highest mean days incarcerated (77 days). The mean discrimination score was highest among Latino/Latina individuals (1.9 out of 24) and the mean identity score was highest among Black individuals (32.9 out of 48).

Table 1

Descriptive characteristics of the National Epidemiologic Survey of Alcohol and Related Conditions-III participants (2012–2013, N = 14,728, weighted N = 635,788,892).

	Race/ethnicity
Characteristics	Blackn = 7,445Weighted n = 26,550,015			Latino/Latinan = 6,804Weighted n = 33,595,359			American Indian/Alaska Nativen = 479Weighted n = 3,433,517			TotalN = 14,728Weighted N = 635,788,892
		Weighted			Weighted			Weighted			Weighted
	n	n	col %	n	n	col %	n	n	col %	N	N	col %
Age group
18–29	1,844	7,049,603	26.5	2,011	10,155,198	30.2	91	652,852	19.0	3,946	17,857,653	28.1
30–39	1,466	4,727,924	17.8	1,723	80,044,498	23.8	95	660,456	19.2	3,284	13,392,878	21.1
40–49	1,471	4,970,152	18.7	1,360	6,591,721	19.6	91	755,639	22.0	2,922	12,317,511	19.4
50+	2,664	9,802,336	36.9	1,710	8,843,943	26.3	202	1,364,571	39.7	4,576	20,020,850	31.4
% Female	4,415	14,603,636	55.0	3,806	16,859,041	50.2	279	2,010,899	58.6	8,500	33,473,576	52.6
Education
Less than high school	1,257	4,261,718	16.1	2,088	10,204,026	30.4	73	524,212	15.3	3,418	14,989,956	23.6
High school completion or GED	2,451	8,452,645	31.8	1,921	9,439,928	28.1	136	887,946	25.9	4,508	18,780,519	29.6
Some college but did not graduate, orreceived an associate’s or technicaldegree	2,600	9,433,710	16.6	1,978	9,463,447	13.4	199	1,425,340	17.4	2,025	20,322,497	14.9
College completion, bachelor’s degreeor higher	1,137	4,401,941	35.5	817	4,487,959	28.2	71	596,020	41.5	4,777	20,322,497	31.9
Lifetime drug or alcohol use disorder (% yes)	1,878	6,784,041	25.5	1,660	8,266,116	24.6	219	1,611,499	46.9	3,757	16,661,656	26.2
% Ever incarcerated	1,235	4,368,515	16.5	771	3,656,413	10.8	135	900,574	26.2	2,141	8,925,502	14.0
Days incarcerated (weighted mean, 95% CLM)		77.5 (64.5, 90.5)			31.4 (23.9, 38.7)			46.4 (26.5, 66.3)			51.5 (44.6, 58.2)
Racial discrimination score (0–24; weighted mean, 95% CLM)		1.1 (1.0, 1.2)			1.9 (1.8, 2.0)			1.1 (0.8, 1.3)			1.5 (1.4. 1.6)
Racial identity score (0–48, weighted mean, 95% CLM)		32.9 (32.6, 33.0)			32.8 (32.6, 33.1)			28.9 (28.1, 29.7)			32.6 (32.5, 32.7)

In our fully adjusted ZIP model, racial identity was positively associated with having zero days incarcerated (β = 0.132, SE = 0.001, p<0.0001), while racial discrimination was negatively associated with having zero days incarcerated (β = -0.199, SE = 0.001, p<0.0001) (Table 2). Likewise, from the linear portion of the model, racial identity was negatively associated with the total number of days incarcerated (β = -0.110, SE = 0.000, p<0.0001), while racial discrimination was positively associated with the total number of days incarcerated (β = 0.063, SE = 0.001, p<0.0001). These trends were mostly consistent between racial/ethnic groups; the association between identity and fewer days incarcerated was strongest among Black respondents (β = -0.147, SE = 0.001, p<0.0001) and the association between discrimination and more days incarcerated was strongest among AI/AN respondents (β = 0.174, SE = 0.000, p<0.0001).

Table 2

Model results for zero-inflated Poisson regression (NESARC-III, 2012–2013).

	Blackn = 7,445			Latino/Latinan = 6,804			American Indian/Alaska Nativen = 479			Totaln = 14,728
Zero Model	β	SE	p	β	SE	p	β	SE	p	β	SE	p
Identity
Tertiles 1–3	0.086	0.001	<0.0001	0.183	0.002	<0.0001	0.305	0.004	<0.0001	0.132	0.001	<0.0001
Discrimination
Tertiles 1–3	-0.202	0.001	<0.0001	-0.262	0.001	<0.0001	-0.213	0.003	<0.0001	-0.199	0.001	<0.0001
Linear model
Identity
Tertiles 1–3	-0.147	0.000	<0.0001	0.001	0.000	<0.0001	0.093	0.000	<0.0001	-0.110	0.000	<0.0001
Discrimination
Tertiles 1–3	0.070	0.000	<0.0001	-0.031	0.000	<0.0001	0.174	0.000	<0.0001	0.063	0.000	<0.0001

We also used output generated by the ZIP models to estimate the predicted days incarcerated across levels of racial identity and discrimination (Fig 1A–1D). Racial discrimination was positively associated with predicted days incarcerated for the total sample and across all three racial/ethnic groups. Changes in days incarcerated were most notable among Black and AI/AN respondents. Black respondents with low discrimination exposure had 42 predicted days incarcerated, whereas Black respondents with high discrimination exposure had 130 predicted days incarcerated, or an increase of 209%. Similarly, AI/AN respondents demonstrated an increase of 106% between low and high levels of discrimination. Racial identity was negatively associated with predicted days incarcerated for the total sample and for all three racial/ethnic groups. The biggest change was observed among Hispanic respondents. Those with low levels of racial identity had 37 predicted days incarcerated, while those with high levels of racial identity had 17 days incarcerated, or a decrease of 54%.

Fig 1

Predicted days incarcerated by racial discrimination and racial identity scores for the total sample and for Black, Latino/Latina, and American Indian/Alaska Native subgroups.

A: Predicted days incarcerated by racial discrimination and racial identity scores (NESARC-III; Total N = 14,728, weighted N = 635,788,892). B: Predicted days incarcerated by racial discrimination and racial identity scores (NESARC-III; Black n = 7,445, weighted n = 26,550,015). C: Predicted days incarcerated by racial discrimination and racial identity scores (NESARC-III; Latino/Latina n = 6,804, weighted n = 33,595,359). D: Predicted days incarcerated by racial discrimination and racial identity scores (NESARC-III; American Indian/Alaska Native n = 479, weighted n = 3,433,517).

Supplemental results

Pearson’s correlation coefficients for all analytic variables are presented in S1 Table. We identified no multicollinearity between predictor variables using a threshold of r<0.80. We also derived the predicted days incarcerated for each individual discrimination and identity survey item, stratified by race/ethnicity and adjusted for age rate, sex, highest grade completed, and alcohol/drug use (S2 Table). We reported the predicted days incarcerated at the lowest response value and the highest response value for each question, as well as the percentage difference between low and high scores. For example, among Black respondents, the predicted days incarcerated for those with no experience with discrimination in healthcare settings was 33.6, whereas the predicted days incarcerated for those that have experienced discrimination in healthcare ‘very often’ was 66.2, or a 97.0% increase in days incarcerated between ‘no’ discrimination and ‘very often’ discrimination.

Discussion

In this study, we examined racial/ethnic differences in associations between incarceration and racial discrimination and identity. Our findings highlight the carceral implications of exposure to racial discrimination while controlling for the effect of racial identity. Higher racial identity scores were associated with fewer days incarcerated, which illustrates a partial buffering effect of identity on discrimination. Days incarcerated were highest among AI/AN and Black individuals with high levels of racial discrimination, and the buffering effect of identity appeared to be the strongest among AI/AN individuals. This is generally aligned with previous research that found that racial discrimination had a weaker effect on depression among Black individuals with strong racial identities [18]. Future research may consider identifying the mechanisms underlying the buffering influence of identity on discrimination.

Our findings are in agreement with previous work showing associations between racism, discrimination, and incarceration [28,29]. Importantly, we have added much needed nuance in our measurement of discrimination and identity, and how these metrics relate to a tangible outcome–the number of days incarcerated. Visualizing these relationships provides the often missing contextual pieces of racism and culture within narratives of criminal justice and public health. While the mean number of days incarcerated was highest among Black respondents (78 days), we also identified that the AI/AN population reported the highest rate (26%) of ever being detained in their lifetime. The impacts of this cannot be understated, especially given the impact that prior incarceration has on job application and employment opportunities [30], discrimination and stigma based on conviction status [31], and adverse experiences among children of incarcerated parents [32]. By directly addressing social determinants of health across multiple racial/ethnic minoritized groups, the current study illustrates a measurable outcome of structural racism in an easily understood and relatable metric of days spent in a jail or prison setting.

Strategies aimed at reducing incarceration should consider the nuances of discrimination and identity, as well as how they differ between racial/ethnic groups. In particular, interventions taking place within lower ecological levels may serve to weaken the effects of discrimination by empowering those who experience it. For example, empowerment-based approaches that support inclusive communities and teach resiliency and coping may help to advance health equity while diminishing the clout of discriminatory individuals and institutions. In addition, targeting higher ecological levels may move the national discourse towards changing social norms and supporting policies, practices, and built environments that engender racial equity.

Limitations

NESARC-III considers lifetime incarceration as ever having been in jail or prison. However, jails and prisons differ in many ways, and racial discrimination and identity may affect the likelihood of incarceration differently between the two locations. Exposure to racial discrimination may also differ within jails versus prisons. Future research may highlight more granular effects by differentiating between jails and prisons.

Respondents self-reported their experiences with ever being detained in a juvenile detention center or jail in their lifetime. People may be reluctant to disclose all or part of their criminal histories to interviewers, resulting in the possibility of introducing desirability bias to the incarceration estimates. Such a bias would likely result in a more conservative underestimation of the reported days incarcerated, and it is unknown whether there is variability in underreporting between racial/ethnic groups. To our knowledge, there exist no other national jail datasets that have sample sizes and complex survey designs analogous to NESARC-III. The Bureau of Justice Statistics (BJS) estimated that 2–3% of the US population had been detained in a jail in 2019 [33]. In comparison, our estimate that 14% of adults reported ever being detained is reasonable, given that lifetime estimates require recalling many more years than were included in the BJS 2019 report [33]. Furthermore, regional and state-level jail datasets mostly report on aggregate annual admissions–national data on lifetime estimates of days detained at the individual level does not exist elsewhere. Our finding that 14% of adults had ever been detained in their lives is reasonable. Thus, NESARC-III is an imperfect but useful tool to estimate trends in incarceration.

Geographic data, including state, is not available in NESARC-III. We were therefore unable to test external, state-level variables that likely affect incarceration risk. Our inability to include metrics of structural racism like residential segregation, racial/ethnic population densities, policies around health and social services, or criminal justice and policing was a limitation of this study.

Finally, the cross-sectional nature of our data makes inferences of causality impossible. Rather, we are limited to conclusions of associations of non-temporal events. Future longitudinal studies would be well positioned to detect clear and causal relationships between structural racism, discrimination, and identity and subsequent incarceration. Importantly, experiences of incarceration may also shape later perceptions of discrimination and identity, further justifying our support for longitudinal study designs.

Conclusion

Racial discrimination and identity varied between races/ethnicities, such that both discrimination and identity were highest among Black individuals. Black individuals also demonstrated the highest mean days incarcerated across the lifespan, which was more than double the next highest group of AIAN individuals. We also found a strong relationship between discrimination score and days incarcerated. Most notably, racial/ethnic minority groups who had large proportions of their members with the highest discrimination score–Black and AIAN–were estimated to be incarcerated for a total of 9–10 weeks throughout their lives.

Policies aimed at reversing the trend of mass incarceration should consider how our carceral systems fit within the wider contexts of historical racism and structural determinants of health. We recommend addressing the challenges of discrimination on multiple ecological levels. Examples include educational programs emphasizing the values of racial and cultural differences, community-level campaigns organizing for equitable access to health and financial services, and electoral support for representatives who campaign on evidence-based criminal justice reform.

Pearson’s correlation coefficient matrix.

(DOCX)

Click here for additional data file.

Predicted days incarcerated by individual discrimination and identity survey items.

(DOCX)

Click here for additional data file.

16 Oct 2021

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Yann Benetreau, PhD

Senior Editor

PLOS ONE

Journal Requirements:

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

1. 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

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

2. Please include a caption for figure 1.

[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: Yes

Reviewer #2: Partly

**********

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

Reviewer #1: Yes

Reviewer #2: No

**********

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: The authors conducted a study of racial discrimination, racial identity, and incarceration risk in a national US sample of Black, Latino/Latina, and American Indian/Alaska Native individuals. Overall, the study is well written, interesting, and will be an important contribution to the literature. The methods are appropriate, thorough, and well-described. I only recommend some minor edits to improve clarity and language.

1. Overall, I recommend not using the term "Blacks", and instead using "Black populations", "Black individuals", or "Black participants".

2. When you discuss the survey procedures and design effects, I'd recommend a sentence on what the clusters and strata represent (like groups of counties, cities, etc.).

3. For the last sentence that "These trends were consistent between racial/ethnic groups, and were strongest among AIAN", I would revise to drop the "and were strongest among AIAN", as it looks like Black and Latinx participants had the strongest association between discrimination and incarceration.

4. Some discussion on how racial identity buffers against internalized racism is recommended.

5. The distinction between jails and prisons noted in the limitations is important; I recommend an additional sentence on how racial discrimination may differ between them.

Reviewer #2: Review of The competing effects of racial discrimination and racial identity on the predicted

number of days incarcerated in the US

General comments

As I read the definition of “discrimination,” the authors’ description of it seems imprecise. As they operationalize it, it is not a measure of discrimination per se, but a measure of “number of types of discrimination ever encountered and recorded in the self-report.” I am not sure that I agree with the decision to disregard the data in the survey on the frequency with which discrimination was encountered. In revision, I think the authors should have to both: a. justify this decision and b. report on analyses in which they created a variable(s) that took account of the frequency data. This comment is based on my belief that people who experience discrimination “all the time” are likely to react in different ways than those who experience it “almost never.” The authors may also want to explore whether one particular type of discrimination is a better predictor. After all, with an N of 14 thousand, such analyses should be possible.

Similar explorations should be made and reported with different ways to define racial identity. Number of types of identification ignores the issue of strength of identity. And it is also of interest to know if one particular kind of identification is more or less associated with incarceration.

Analytically, if the dependent variable is number of days of incarceration, then instead of treating age as a confounder, it should be used to create a rate dependent variable that is defined as “number of days incarcerated divided by numbers of years at risk of being incarcerated.” This is because I expect that self-reports of discrimination; racial identity; and number of days incarcerated are all correlated with age. This is certainly true for days of incarceration. The other variables are likely related because of one or both of the following: a. generations have different experiences and beliefs (see Mannheim’s analysis of generations) and/or b. life stage factors. This seems to me to be a very serious error.

I am skeptical of the unvalidated claim that 86% of the sample reported zero days of incarceration. First, these are self-reported data, and people may be reluctant to disclose criminal histories. Second, even if they are willing to disclose, they might minimize it; or they might say “never incarcerated” in order to avoid follow up questions (depending on skip patterns) and shorten the interview. To be credible, this figure should be supported by comparisons with other studies (such as other surveys) that do not focus on alcoholism but perhaps on criminal victimization and other surveys like some NORC has done.

In an analysis such as this, reporting on the correlations among the main analytic data (probably for total, and within racial categories) is needed. Particularly concerning is the possibility that racial identity and racial discrimination experiences could conceivably be very high.

I think that the Conclusions should include a strong statement about the much higher rates of “ever incarcerated” among AIAN. Given the many limitations that accrue to simply reporting that you were ever arrested in job applications and other critical processes, this is of enormous impact.

Discussion and Conclusions discuss structural racism a lot. This makes the analytic failure to include any measures of structural racism stand out as a major defect and limitation. (See below under “important additional comments.”

Important additional comments

1. Perhaps the title should indicate the restricted nature of the sample that was analyzed by adding “among Black, Latino/Latina, and AIAN individuals.” Although the exclusion of whites is not problematic given the theoretical framing and the realities of oppression, the exclusion of Asian Americans is an important limitation and should be noted in the title (I think) and should be justified in Methods.

2. Limitations should include the inability to analyze how locality or workplace influenced incarceration. Measures of structural racism like racial/ethnic population densities, racial/ethnic residential and occupational segregation are almost certainly associated with probability of getting arrested, and likely with both racial discrimination and racial identity.

Minor Comments

1. line 2 “bared” is misleading. Probably “borne” would be best

2. line 5 – 6 “remain…remains” I will stop text editing at this point. It is clear that serious text editing is needed.

3. l 9 I am not sure “disenfranchisement” is the best term for this.

4. p. 5 l 6 – 7: Here or in Methods, the variable that was used to limit the sample to Black, Latino/Latina, and AIAN individuals should be described.

5. Table 1 and first paragraph of Results: Important to state whether mean days incarcerated is for the total analytic sample or only for the subset who reported non-zero on this variable.

6. p. 13 -14: I am not sure that the claim about structural racism in the following sentence is justified by the data (although I think it fully justified in reality): “This study supports the notion that incarceration is driven, in part, by the mechanisms of structural racism and personal discriminatory acts towards racial/ethnic minorities.” I say this because I see no measures of structural racism in the set of independent variables.

**********

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.]

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 PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: Review of The competing effects of racial discrimination and racial identity on the predicted number of days incarcerated in the US.docx

Click here for additional data file.

9 Feb 2022

We thank the journal editors and our two reviewers for the opportunity to revise our manuscript. We received invaluable insight from our reviewers about study design, more accurately defining discrimination and identity, and further exploring the influence of discrimination on incarceration risk. We believe our manuscript has been substantially improved as a result of the thoughtful consideration put into each review and is now suitable for publication in PLOS One. We have responded to each comment in detail below, and have submitted a revised manuscript with changes tracked in MS Word.

Reviewer 1

General comments

1. The authors conducted a study of racial discrimination, racial identity, and incarceration risk in a national US sample of Black, Latino/Latina, and American Indian/Alaska Native individuals. Overall, the study is well written, interesting, and will be an important contribution to the literature. The methods are appropriate, thorough, and well-described. I only recommend some minor edits to improve clarity and language.

Response: Thank you for your critical feedback and positive support for our research. We have responded to each of your comments below.

2. Overall, I recommend not using the term "Blacks", and instead using "Black populations", "Black individuals", or "Black participants".

Response: We appreciate this feedback as we are constantly striving to be conscious of the language we use to describe disparities. We have gone through the manuscript and edited how we describe groups, in line with using “Black individuals” or “American Indian/Alaska Native respondents”.

Methods

3. When you discuss the survey procedures and design effects, I'd recommend a sentence on what the clusters and strata represent (like groups of counties, cities, etc.).

Response: This is a good opportunity to provide more information about the survey design to our readers. The probability sampling procedure of NESARC-III is complex, including primary sampling units, secondary sampling units, segments, dwelling units, and respondents. We added two sentences in our methods section outlining the purpose and process of oversampling in high-minority areas, and direct the readers to the NESARC technical documentation for further information.

Results

4. For the last sentence that "These trends were consistent between racial/ethnic groups, and were strongest among AIAN", I would revise to drop the "and were strongest among AIAN", as it looks like Black and Latinx participants had the strongest association between discrimination and incarceration.

Response: Please see our response to Comments #7 and #8. Our definitions for discrimination and identity have changed to include the full information available on the frequency of discrimination exposure and the magnitude of agreement with identity statements. As such, our regression estimates have changed moderately and our results section has been updated, including the sentence described in the comment above.

Discussion

5. Some discussion on how racial identity buffers against internalized racism is recommended.

Response: Better understanding the mechanisms underlying the buffering characteristics of racial identity is an important future research question that came out of this current study. For this study, we believe that speculating on how identity may be operating to suppress the effect of discrimination exposure is beyond the scope of what we measured. However, we did add several sentences to our discussion section outlining how our findings align with previous research addressing the intersection of identity and discrimination.

Limitations

6. The distinction between jails and prisons noted in the limitations is important; I recommend an additional sentence on how racial discrimination may differ between them.

Response: We have added a sentence about the importance of considering how racial discrimination may vary between the two locales.

Reviewer 2

General comments

7. As I read the definition of “discrimination,” the authors’ description of it seems imprecise. As they operationalize it, it is not a measure of discrimination per se, but a measure of “number of types of discrimination ever encountered and recorded in the self-report.” I am not sure that I agree with the decision to disregard the data in the survey on the frequency with which discrimination was encountered. In revision, I think the authors should have to both: a. justify this decision and b. report on analyses in which they created a variable(s) that took account of the frequency data. This comment is based on my belief that people who experience discrimination “all the time” are likely to react in different ways than those who experience it “almost never.”

Response: Thank you for this very helpful comment. We agree that our original definition of discrimination was imprecise – our single category of ‘any’ discrimination included a wide range of experiences, from almost never to all the time. This variable failed to capture what we are trying to measure, which is the frequency, or magnitude, of exposure to discriminatory behavior.

We replaced the original ‘any/none’ variable with a new variable that takes into account the frequency of discrimination exposure. The new composite scale is based on six survey questions, each with response options of 0 (never) to 4 (very often), resulting in a single scale of 0 to 24. Scores at the higher end of the spectrum result from multiple responses endorsing often or very often exposure. This new approach allows for more granularity in measuring discrimination, as higher scores reflect a greater frequency of discrimination.

8. Similar explorations should be made and reported with different ways to define racial identity. Number of types of identification ignores the issue of strength of identity.

Response: Our original racial/ethnic identity scale was similar to the original discrimination scale, in that we defined ‘any’ positive identity as any respondent that somewhat agreed, agreed, or strongly agreed with each statement. The original variable was based on eight survey questions and the scale ranged from 0 to 8. However, applying the same logic as the reviewer raised above in Comment #7, experiences of individuals who ‘somewhat agree’ are likely different than those who ‘strongly agree’. Therefore we reconstructed our identity variable to reflect the magnitude, or strength, of affirming responses. The revised identity scale now ranges from 0 to 48, with higher scores reflecting a greater magnitude of agreement with each question.

9. The authors may also want to explore whether one particular type of discrimination is a better predictor. After all, with an N of 14 thousand, such analyses should be possible.

And it is also of interest to know if one particular kind of identification is more or less associated with incarceration.

Response: This is a great point, as our composite discrimination and identity scales do not capture the influence of each individual sub-item on incarceration. We created a Supplemental Table 1, which outlines the relationship between each survey item and incarceration. Specifically, we derived the predicted days incarcerated for the lowest and highest scores for each question, and also reported the percentage difference in days. For example, among Black respondents, the predicted days incarcerated for those with no experience with discrimination in healthcare settings was 33.6, whereas the predicted days incarcerated for those that have experienced discrimination in healthcare ‘very often’ was 66.2, or a 97.0% increase in days incarcerated between ‘no’ discrimination and ‘very often’ discrimination.

Title

10. Perhaps the title should indicate the restricted nature of the sample that was analyzed by adding “among Black, Latino/Latina, and AIAN individuals.” Although the exclusion of whites is not problematic given the theoretical framing and the realities of oppression, the exclusion of Asian Americans is an important limitation and should be noted in the title (I think) and should be justified in Methods.

Response: We have changed our title to state specifically the groups included in this study, and have added text in the methods section about how we arrived at our final analytic sample. Our new title is “The competing effects of racial discrimination and racial identity on the predicted number of days incarcerated in the US: A national profile of Black, Latino/Latina, and American Indian/Alaska Native populations”.

Introduction

11. Line 2 “bared” is misleading. Probably “borne” would be best

Response: We have changed ‘bared’ to ‘borne’.

12. Line 5 – 6 “remain…remains” I will stop text editing at this point. It is clear that serious text editing is needed.

Response: We appreciate the opportunity to make our writing as clear as possible for our readers. In this particular case, we have respectfully kept the word ‘remain’, as it refers to the plural noun ‘populations’. We have reviewed our manuscript thoroughly for grammatical errors and typos, and have edited sentence structure to be as clear and correct as possible.

13. Line 9 I am not sure “disenfranchisement” is the best term for this.

Response: We have changed ‘one form of disenfranchisement’ to ‘one factor in particular’.

Methods

14. Analytically, if the dependent variable is number of days of incarceration, then instead of treating age as a confounder, it should be used to create a rate dependent variable that is defined as “number of days incarcerated divided by numbers of years at risk of being incarcerated.” This is because I expect that self-reports of discrimination; racial identity; and number of days incarcerated are all correlated with age. This is certainly true for days of incarceration. The other variables are likely related because of one or both of the following: a. generations have different experiences and beliefs (see Mannheim’s analysis of generations) and/or b. life stage factors. This seems to me to be a very serious error.

Response: This is a great point and a methodological oversight on our part. While we still report on age groups in our descriptive Table 1, we have replaced age group with a rate dependent variable in our predictive models.

15. I am skeptical of the unvalidated claim that 86% of the sample reported zero days of incarceration. First, these are self-reported data, and people may be reluctant to disclose criminal histories. Second, even if they are willing to disclose, they might minimize it; or they might say “never incarcerated” in order to avoid follow up questions (depending on skip patterns) and shorten the interview. To be credible, this figure should be supported by comparisons with other studies (such as other surveys) that do not focus on alcoholism but perhaps on criminal victimization and other surveys like some NORC has done.

Response: Recognizing possible bias in responses to sensitive questions about incarceration history is critical for this study, and not including this as a limitation was an oversight. We have included a paragraph in our limitations section outlining how the possibility of desirability bias could affect the reported estimates of incarceration. In addition, we compared our findings to BJS jail admission data, and concluded that while NESARC-III is imperfect, it is a useful and validated tool to estimate population characteristics, including experiences of incarceration.

16. In an analysis such as this, reporting on the correlations among the main analytic data (probably for total, and within racial categories) is needed. Particularly concerning is the possibility that racial identity and racial discrimination experiences could conceivably be very high.

Response: This was helpful for us to think through the complex relationships between all of our analytic variables. We created a new Supplemental Table 2 which includes all of the information from a Pearson’s correlation matrix, for the total sample and stratified by race/ethnicity.

17. The variable that was used to limit the sample to Black, Latino/Latina, and AIAN individuals should be described.

Response: We have expanded our description of the race/ethnicity variable we used. NESARC=III provides a single variable that includes pre-constructed categories of race and ethnicity. Please also see our response to Comment #10, where we expanded our description of each racial/ethnic category available in NESARC-III.

Results

18. Table 1 and first paragraph of Results: Important to state whether mean days incarcerated is for the total analytic sample or only for the subset who reported non-zero on this variable.

Response: We appreciate the opportunity to clarify how we are reporting the study results. The mean days incarcerated is for the total sample, and we have clarified this accordingly in the Results section.

Discussion

19. I think that the Conclusions should include a strong statement about the much higher rates of “ever incarcerated” among AIAN. Given the many limitations that accrue to simply reporting that you were ever arrested in job applications and other critical processes, this is of enormous impact.

Response: Thank you for this comment – we had not discussed this impactful finding in our original submission but we agree that it deserves attention in our revised discussion section. We have added several sentences and three new citations summarizing the impact of such high rates of incarceration among American Indian/Alaska Native respondents. We also brought to our reader’s attention the juxtaposition of Black respondents having the highest mean number of days incarcerated (78 days), while AI/AN respondents had the highest proportion of any lifetime incarceration (26%).

20. Discussion and Conclusions discuss structural racism a lot. This makes the analytic failure to include any measures of structural racism stand out as a major defect and limitation.

Response: We value the opportunity to clarify for our readers how a discussion of structural racism fits in to the overall narrative of race, discrimination, and incarceration. We could not identify any variables within the NESARC dataset that closely resembled anything like structural racism. NESARC does not include a variable for state, which would allow researchers to identify other external state-level data that would serve as a proxy indicator of system-level discrimination and racism, like residential segregation, health provider shortage areas, or state-wide financial estimates of disproportionate loan/mortgage rejections among poor and racial minority groups, for example. At the same time, we believe that incarceration itself is a form of structural racism. We therefore feel that it is our responsibility to engage in a conversation about structural racism, and that this conversation is within the scope of our findings about racial discrimination and incarceration. While not directly measured or tested in this study, structural racism is the language that we choose to use when contextualizing the mass incarceration epidemic.

21. I am not sure that the claim about structural racism in the following sentence is justified by the data (although I think it fully justified in reality): “This study supports the notion that incarceration is driven, in part, by the mechanisms of structural racism and personal discriminatory acts towards racial/ethnic minorities.” I say this because I see no measures of structural racism in the set of independent variables.

Response: We agree that this sentence in particular is worded in a way that one may assume we directly tested indicators of structural racism. Even in light of our defense of our discussion around structural racism (see Comment #20), we have removed this sentence to make our study design and focal independent variables as clear as possible.

Limitations

22. Limitations should include the inability to analyze how locality or workplace influenced incarceration. Measures of structural racism like racial/ethnic population densities, racial/ethnic residential and occupational segregation are almost certainly associated with probability of getting arrested, and likely with both racial discrimination and racial identity.

Response: We have included a new paragraph in our Limitations section that brings to attention the lack of geographic data in NESARC-III, including state. At the very least, access to data about which state the survey respondent resides in would allow for the use of external, state-level indicators of social, environmental, occupational, and economic determinants of health.

Submitted filename: PLOS discrimination_Response to reviewers.docx

Click here for additional data file.

21 Mar 2022

28 Mar 2022

Comments from the editor

ABSTRACT

Comment 1: Page 2, Lines 11-12: change ‘discrimination exposure’ to ‘racial discrimination exposure’.

Response: We have made this change.

Comment 2: Page 2, Line 15: What AI/IN stand for? please spell out these acronyms and other acronyms on their first appearance.

Response: Thank you for catching this. We have spelled out the acronym for American Indian/Alaska Native at first use, which is in the second sentence of the abstract.

Comment 3: Page 2, Line 19: The conclusion should be based on / refer to the results, which show that Racial discrimination and racial identity are associated with incarceration.

Response: Thank you for helping us better organize our abstract conclusion. We have reframed this paragraph to restate our findings about discrimination and identity. At the same time, we respectfully feel that it is within the scope of this study to reference mass incarceration and racism as drivers of disparities. Later in the paper, our discussion uses these concepts to contextualize increased risk for involvement in the criminal justice system among underrepresented groups. As an epidemiologic study through a public health lens, we believe that failing to articulate the upstream factors affecting incarceration would be a disservice to our audience. In addition, please see Comment 12 below (Reviewer 2). The sentiment of this feedback was that the abstract conclusion should focus more on these same issues we raised in the discussion section, including structural determinants of health and their effect on incarceration and community health.

INTRODUCTION

Comment 4: Page 3: Please revise ‘upwards of 10 times higher’. It should be either ‘upwards’ or ‘higher’.

Response: We have rephrased this sentence to read, “nearly 10 times higher”

Comment 5: Page3, lines 21-22: Please provide a citation to support the statement: Eighty-five percent of Blacks and 51% of Latino/Latinas study participants in California reported ever being treated unfairly because of their race/ethnicity, compared of Whites participants.’

Response: We have removed this statement.

METHODS

Comment 6: Page7, lines-9-10: Please check the statement: ‘educational attainment (less than high school, high school completion, some college, or college graduate), what’s the difference between ‘some college’ and or ‘college graduate’? These could be merged in one category.

Response: Thank you for the opportunity to clarify our methods and variable creation. We agree that there are multiple ways to combine and rearrange this education variable. In this case, there is an important distinction between ‘some college’ and ‘college graduate’, such that the former refers to some experience with college coursework but did not finish, while the latter refers to those who graduated with a bachelor’s degree. College completion in particular has multiple implications for employment prospects and income, and may be a proxy indicator for overall economic stability. For these reasons, we have respectfully maintained the separation between ‘some college’ and ‘college graduate’.

Comment 7: Page 7, lines11-12: Please refer to your statement: “We also included a rate dependent age variable, defined as the number of days incarcerated divided by the number of years at risk of being incarcerated.” What is 'rate dependent age variable? It could be rate of something depending on age. Based on the definition given in the next sentence, it could be named as 'incarceration age rate' or 'incarceration risk rate'.

Response: This is a great question and we appreciate the chance to clarify the definition of all of our study variables. We added this variable to our analysis in response to our second round of peer review. One reviewer specifically suggested that we include this variable, named ‘rate dependent age variable’, and we defined it exactly as recommended by the reviewer. Given that both reviewers have provided two separate rounds of feedback, and both are currently satisfied with the methods and variables as they are (below), we are reluctant to make any further changes that may contradict our response to their previous review.

LIMITATIONS

Comment 8: Page 15, line13: BJS 2019 report. Could you please spell out what BJS stands for? Also provide a citation for the report.

Response: We have included the acronym ‘BJS’ at the first use of the phrase ‘Bureau of Justice Statistics’. We have also made it clear that the citation for the BJS report (included in two sentences) is for citation #33, or Zeng et al.

REFERENCES:

Comment 9: Please report journal names in the abbreviated form, where available.

Response: We have changed the journal names to the abbreviated form.

Reviewer #1

Comment 10: The authors have successfully addressed my concerns. I have no further recommendations regarding the manuscript.

Response: Thank you for your thoughtful and careful consideration of our manuscript.

Reviewer #2

Comment 11: This is an excellent revision and an important paper.

Response: Thank you for your time and energy put into this critical peer review.

Comment 12: I have one suggestion about how to improve the Abstract, though I view it as optional for the authors. As currently written, the Conclusions section of the abstract seems so general as to convey little meaning. I think the last paragraph of conclusions in the text has excellent suggestions for interventions, and these should be the focus of the conclusions in abstract. This will also, in my opinion, lead to more people reading the paper.

Response: Thank you for your positive feedback about our discussion of structural drivers of inequity. Please see our response to Comment 3 by the journal editor. We have kept most of the language used to contextualize our findings and reflect the points we raise in the discussion section about systemic racism and discrimination. At the same time, we are striving to balance this contextual narrative with restating the study findings and keeping the abstract as succinct as possible, as per the journal editor’s comment referenced above.

Submitted filename: PLOS response to reviewers, round 3.docx

Click here for additional data file.

9 May 2022

10 May 2022

Comments from the editor

Comment #1: In your last reply to the AE/Reviewers, you have stated that 'some college' category of the educational attainment refers to those who had "some experience with college coursework but did not finish". If so then this category could be better reported as 'college student', if studying at the time of completing the survey or 'college dropouts' if they had left the college without completing/graduating. Please address this issue in the text as well as Tables including supplemental material.

Response: We appreciate the opportunity to clarify how this variable was created and defined. Please see our response to Comment #4. We have added several sentences to the methods section more clearly outline the specific survey response options that were included in our condensed analytic variable. We have also changed the names of the variable levels to more accurately reflect the highest level of education attained by each survey respondent. We updated Table 1 with the new labels. Table 2 and the supplemental material did not reference the education variable levels.

Comment #2: Tables 1 and 2: Please shift the column 'Total' as the last column on the right hand side because the focus of the study is on the ethnic/racial groups and not on the total/aggregate of the groups.

Response: We have shifted the total column to the right hand side for both Table 1 and Table 2.

Comment #3: Three categories of tertiles: low, mid, and high have been used/reported in the text, table 2 and figures 1A-D. Could you please report in the methods section, how low, mid and high tertiles were determined and what values are covered by each of these three tertile categories?

Response: We have added a paragraph in the methods section defining how tertiles are created based on the distributions of the identity and discrimination scores around the 33rd and 66th percentiles. We have also included the range of identity and discrimination scores for each of the tertile groups.

Reviewer 2

Comment #4: I have one minor and optional suggestion for a wording change: When you define the variable about education, the precise definition of "some college" and "college graduate" remains unclear. The issues are that one can graduate from a junior college after 2 years; and that college graduate also have some college. So it might be useful to say what you said in the response note, that college grads means undergraduate degree.

Response: We appreciate the opportunity to clarify how we are defining education, as we agree that the current wording is confusing. While education is not a focus of this study, it acts as an important confounder in our models and we believe it is important for readers to understand how this variable was constructed. We have added text in the methods section defining each of the survey options that we combined to create the new condensed analytic variable. We have also edited the labels for each level to more clearly articulate the highest level of education attained. The new labels are 1) less than high school, 2) high school completion or GED, 3) some college but did not graduate, or received an associate’s or technical degree, and 4) college completion, bachelor’s degree or higher.

Reviewer 3

Comment #5: Page 7, line 18, “We grouped discrimination scores (0-6) and identity scores (0-8)” should be revised as “We grouped discrimination scores (0-24) and identity scores (0-48)”.

Response: We have corrected this typo.

Comment #6: Page 13, Supplemental Tables 1 and 2 are mentioned, but nowhere can find them.

Response: Supplemental tables 1 and 2 were included as attachments in the PLOS online application portal. We followed specific instructions not to embed supplemental material in the main document text. I can confirm that they are also included as attachments in this current resubmission. If the supplemental tables do not appear on your end again, please reach out to the academic editor.

Comment #7: Are there any correlations and/or interactions between racial discrimination and racial identity?

Response: This is a great question and we appreciate your interest in our research. We believe that testing for correlation or an interaction between identity and discrimination is beyond the scope of the current study, but is very much of interest to us for future research endeavors. For this study, we are specifically interested in the main effects of identity and discrimination on incarceration in a fully adjusted model. Testing the hypothesis that the effect of identity on incarceration is conditional on the value of discrimination would be a valuable contribution to the social sciences and criminal justice literature. At the same time and respectfully, we have opted to make no changes to our statistical model and consider your question about alternative methods as positive motivation to continue this line of research.

Submitted filename: PLOS response to reviewers, round 4.docx

Click here for additional data file.

13 May 2022

The competing effects of racial discrimination and racial identity on the predicted number of days incarcerated in the US: A national profile of Black, Latino/Latina, and American Indian/Alaska Native populations

PONE-D-21-10832R3

Dear Dr. Pro,

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 authorbilling@plos.org.

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 onepress@plos.org.

Kind regards,

Syed Ghulam Sarwar Shah, M.B.B.S., M.A., M.Sc., Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Thanks for addressing issues raised by the academic editor and reviewers.

However, I am not aware of the US educational system so I am not sure whether "attended but did not finish graduate school' is correct in the following statement (lines 14-16 on page 7 of article file with track changes). This statement needs to be checked at the article proof checking stage.

"College graduates were defined as those who received a bachelor’s degree, attended but did not finish graduate school, or completed a master’s degree or equivalent or another higher graduate degree."

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 #2: All comments have been addressed

Reviewer #3: 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 #2: (No Response)

Reviewer #3: (No Response)

**********

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

Reviewer #2: (No Response)

Reviewer #3: (No Response)

**********

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 #2: (No Response)

Reviewer #3: (No Response)

**********

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 #2: (No Response)

Reviewer #3: (No Response)

**********

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 #2: (No Response)

Reviewer #3: (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 #2: Yes: Samuel R Friedman

Reviewer #3: No

18 May 2022

PONE-D-21-10832R3

The competing effects of racial discrimination and racial identity on the predicted number of days incarcerated in the US: A national profile of Black, Latino/Latina, and American Indian/Alaska Native populations

Dear Dr. Pro:

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 onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

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

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Syed Ghulam Sarwar Shah

Academic Editor

PLOS ONE