Do Segmented Assimilation Theory and Racialized Place Inequality Framework Help Explain Differences in Deaths Due to COVID-19 Observed among Hispanic Subgroups in New York City?

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The study uses rigorous methodology to account for death rates due to Covid- 19 among different Hispanic subgroups in NYC. The manuscript takes into account the spatial clusters as well as various other socio economic factors among Hispanic subgroups. 

 

The study’s strength lies in developing a theoretical model and its attempt to explain the findings through further exploratory analysis.  Comments:

 

1. The overall structure of the manuscript could be improved to make the findings of the main model and further exploratory analysis more clear. It will help the readers if the key findings are clearly stated.
2. The study should have a set of research questions and testable hypotheses in the beginning, as well as a  clear motivation.
3. The importance of spatial factors need to be more highlighted in light of the Covid 19 situation. What is the role played by spatial clusters in the death rates by Hispanic subgroups?
4. The difference between different Hispanic subgroups should be clearly stated and the motivation for studying Covid -19 impacts by these subgroups should be explained. How does it make a difference where their origin was? Is there a different pattern of death found across different Hispanic subgroups that calls for this analysis?
5. The paper has good information which needs to be restructured in a more legible and crisp manner.  The number of tables can be reduced and the results of further exploration can be put in the discussion section and in the appendix. 

Author Response

ANSWERS

1. The overall structure of the manuscript...

Answer: Thank you for the comment. The structure has been modified to motivate the research as showing evidence of the different assimilation paths for immigrants that are proposed by Segmented Assimilation Theory and RPIF. Findings are also now clearly stated. The entire paper has changed to address your concerns.

2. The study should have a set of research questions...

Answer: Thank you for the comment. The study now has three research questions that are shown in the introduction.

3. The importance of spatial factors...

Answer: Thank you for the comment. The study now motivates theoretically the importance of spatial correlation between lines 75 and 97.

4. The difference between different Hispanic subgroups.

Answer: The importance of the different Hispanic subgroups and a theoretically motivated argument for such differences to exist is now discussed between lines 44-74.

5. The paper has good information:

Answer: The paper has been rewritten. New maps are included in section 2.1, some tables have been dropped from the text, while all tables that show results are now discussed in terms of its relevance to show the answers to the different research questions.

Reviewer 2 Report

Comments and Suggestions for Authors

This paper explores how risk and protective factors affect death due to COVID for different Latino groups in NYC. The analyses are conducted at the PUMA level. Overall, the author(s) provide correlational analyses of which factors increase or decrease death rates. They use the risk-factor approach described by House (2002)—see reference below—and present many tables to provide detailed analyses. The discussion and conclusion are well done, and the author(s) extrapolate possibilities that might explain the patterns they document. The propositions they outline set the stage for future research.

Below, I provide extensive comments and critique the paper with an eye on how to improve the manuscript. After my observations, I provide my view on how to address the issue with an asterisks * . I begin with comments about different sections (bolded and italicize by heading)   and I conclude with important framing issues.  

In the Introduction

*Briefly explain what the Hispanic Paradox is. You must assume that the reader is not familiar with this term. You can simply state that although research has shown that lower SES people have worse health, first generation lower SES Hispanics are healthier. This is why it is a paradox. You do this further down the paper, but briefly explain it in the introduction.

*Break the Hispanic Paradox down to the reader.

You argue that “Finding a positive correlation between the Hispanic share and the death rate is surprising given that previous research has documented the existence of a higher life expectancy among the Hispanic population”.

Be careful with this statement because this only applies to the first generation and the 1.5 generation. After that, a higher share of the Hispanic population is correlated with poorer health.

*Be more precise with that statement in terms of generational and immigration status. The way it reads is inaccurate.

What is your unit of analysis for this statement, cities, counties, or individuals? The PUMA?

*Please clarify the unit of analysis.

In the introduction, you list several risk and protective factors: doctor density, the length of stay in the US, the citizenship status, the access to employment, the access to health insurance, the civil status, and the headship condition.

House (2002) makes a convincing case for why the risk factor approach is not ideal. It leads to atheoretical work and a variable approach to the literature. House’s main issue with the risk factor approach is that the body of literature then is based on an infinite number of variables with no conceptual and theoretical coherence.

*You should consider abandoning the risk-factor framing of the paper in the front end. Instead, you should frame it conceptually and theoretically to address a problematic in the literature. For instance, you could argue that most studies have looked at individual level characteristics (i.e., risk factors). Few studies, if any, have looked at PUMAS as a macro-level structure. Place stratification scholars (see Velez 2017, Burgos, Rivera, and Garcia 2017) in sociology, and others in the less theoretically developed epidemiological literature (Krieger 2021), have argued that health scholars need to take context seriously. And, they should contextualize context within a theoretical framework.

This is a nice problematic for the framing of the paper because it gives you a justification for looking a PUMA’s as a macro-level social structure. The rest of the paper would flow from there. PUMAS matter because they are a significant geographical unit. Make the case why this is so within a theoretical framework.

In your different sections of the literature review

1.1  Hispanic Paradox

You mention many of the risk factors and protective factors of Latino health. But in your measures, you don’t include many of these factors. The ones that you include come out of nowhere when you finally introduce them.

*The fix: cite literature in the front end that highlight the risk factors that you use in this paper. Then make the case that although the risk factors we use have been identified in the literature, they have not been explored among different subgroups. No study has examined how these risk-factors operate at the PUMA level. This is important because PUMAS are important macro-structural units.

*Most importantly, separate risk factors at the PUMA and the individual level and discuss them separately in the literature review. At this point of the paper, it is not clear that you are doing a PUMA level analysis. Make this clear. Again, you can also highlight that PUMAS, as a macro-level structure has not been examined with these risk factors.

*Make the case for why PUMAs matter. One possibility is that PUMAS are less of a distal social structure than cities and metropolitan areas.

*Throughout the paper, make sure, that you do not equate PUMAS with neighborhoods. More on this below. PUMAs are not neighborhoods nor do can they be used for proxy’s of neighborhoods. Clean up the neighborhood language by not alluding that PUMAS are a proxy for neighborhoods.

1.2 Structural Racism:

*You should cite the ASR article by Bonilla Silva on structural racism and his work, in addition to the references that I include below.

*You talk about implicit racism and define it. What you are talking about is implicit or subtle discrimination. Do not use the word implicit racism. Use the word implicit discrimination or subtle discrimination. It is very important to make this distinction.

Racism is structural and you can have racism without racist. Discrimination is measured at the Individual level. To make this distinction clear you need:

A conceptual model to frame the discussion, such as the Racialized Place Inequality Framework (RPIF) (see Velez 2017; Burgos Rivera (2017) or Ecological Theory (see Krieger 2021). The RPIF is more appropriate given your discussion.

*Frame your literature review within the RPIF and then it becomes easy to argue that PUMAS are an important social structure. You can add a small section after the introduction called theory. Witin the RPIF, you can easily differentiate between racism and discrimination.

*Importantly, you don’t have a measure of structural racism. I’m not sure why structural racism is part of your framing—more on this below. Get rid of mentions of structural racism in the front end and move such a mention to the discussion; preferably when discussing future research. This would be an important insight and you can work on other articles to follow-up on the effects of structural racism on COVID.

Section 1.4 Spatial Correlation

*Please provide a summary sentence or two of what specific conditions (i.e., variables) explain the spatial correlation. Make it clear what spatial correlation is to the reader who know nothing about spatial correlation. Below, in your results section, you also need to do a better job explaining what direct, indirect, and total effects represent within the context of spatial correlation.

*In all these sections, identify the relevance of the findings for your current paper. How do you move the literature forward? What is your contribution? In general, how does the section relate to your current work. Remind the reader as you go through these sections what research question is answered by that section. Otherwise, it reads like a list of things unrelated to the current paper.

Section 1.5 Heterogenity.

*Mention the fact that Latinos are an ethnically heterogeneous population. Give a citation indicating why it’s important to capture this heterogeneity. Google the term “Ethnic Lumping” and Latino and you should get the appropriate reference. Then, highlight why it’s important to capture heterogeneity. Cite the Fuentes and Burgos (2017) article in this section, and which groups have worse health outcomes.

*You discuss the small literature review in the 1.5 section article by article. Instead, extrapolate patterns that emerge from this literature and discuss these patterns. For example (pattern 1, Citations 1-5, pattern 2, citations 6-7, etc). The topic sentence should be the pattern rather than what the article said.  The article-by-article review is unconventional and one that I tell graduate students not to do. These are just a few hand-picked examples to support the heterogeity argument. What you really want to do in this section are two things: 1.) What does the literature say in terms of patterns, 2) what does the literature say in relation to your question. Remind the reader what is your contribution in a concluding sentence or two for this section.

*There are many other scholars talking about heterogeneity aside from Garcia et al. Cite someone else as well.

*Importantly, relate this section to your research question and what you contribute to this literature.

2. Materials

*Please get rid of the sentence stating that you are using PUMAS to approximate neighborhoods. You cannot equate a PUMA with a neighborhood. If you draw on the Racialized Place Inequality Framework (see Burgos, Rivera, Garcia 2017), for instance, you can make the case that a PUMA is a macro-level social structure. And, that scholars have argued that we need to look at how macro-structures impact health. You don’t need to frame the PUMA as a neighborhood to make this study relevant. Your argument is about context, structure and space, not neighborhoods.

*Please indicate your overall sample size, and sample size by Puma.

* How many PUMAS are in your study.

* Is there spatial autocorrelation if you just include your Dependent Variable? Is this spatial auto-correlation statistically significant.

*Include a map of NYC PUMAS to orient the reader. You can include that as a Figure 1. The US Census provides these maps.

*What software package and version did you use to create the dependent variable and coduct the analyses? This is important for replicability.

*You mention several control variables. Are these variables aggregated at the PUMA level? If so, you are conducting a PUMA level analysis. It is not clear to me at this point o the paper. Please clarify and specify.

*You mentioned that the control variables are linked to structural racism. Everything is, really. You need to specify how and why in a sentence or two. As I read your paper, especially your framing at the front end, I thought that you were going to explore the effects of structural racism. You don’t measure structural racism so please don’t make this term part of the framing. You should move this term to the conclusion and argue that scholars should look at the effects of structural racism, measured at the PUMA level, or at other macro-level geographical units.

*You mention that spatial correlation is more accurate than OLS. How and provide a citation.

*You mention that spatial models allow you to explore direct and indirect effects. This comes out of nowhere. At the very least, these important indirect and direct effects should be part of the framing at the front end. Tell the reader what a direct, indirect, and total effect is in relation to the research question.

3.1 Data Sources

*Data sources should not be placed under Results. It should be placed under the previous section, materials.

When you say the data from zip codes on mortality was cross-walked to PUMAS, do you mean you are aggregating the zipcode data to PUMA.

*Please provide more clarification for the sake of replicability.

As I read on, it is clear that you are aggregating the data the PUMA level.

*Also clarify this above. You need to state this more clearly in the front end.

*The abstract should say a PUMA level analyses so that the unit of analyses is clear from the onset.

*Puerto Ricans make up about 32 percent of the NYC population. Why are they omitted? This needs clear justification. The size of the Puerto Rican population fits your group selection criteria, yet, Puerto Ricans are omitted.

Table 2. You provide descriptive statistics for all your variables.

*You should also provide a Table that compares mean differences in these variables by Latino ethnicity. After all, you are making the central argument that ethnic heterogeneity matters. This is your chance to document the bivariate relationships between all your variables by ethnicity. Are there significant differences by ethnicity? This is an important first step. Again, this is a central issue in your framing and there are no analyses on Latino Ethnic differences in all these risk-factors. Also, do the same for gender difference across all these variables and test for bivariate significance. A wald test, a t-test, a chi-square test should be used where appropriate. Only then can you move to Table 3.

When describing Table 3, direct and indirect effects, you need to assume that the reader is not familiar with this technique.

*Begin by describing what research question Table 3 explores. What is a direct and indirect effect. What do these things mean. This will orient the reader to the table and help them follow your explanation.

When describing Table 3, you state, when introducing controls: “The positive

correlation vanishes for US born Hispanics, which implies that structural racism and spatial correlation explained the previous positive result”.

*You can’t say this because you do not measure structural racism. All you have are various controls for occupations, health insurance, and unemployment. At the very least, when you describe your variables, you should have a structural racism paragraph. What it is, how it’s measured in the literature, and how you measure structural racism. None of your variables measure structural racisms. See citations below on how to do this.

*This is an easy fix: Do not argue that you are measuring structural racism and that structural racism explains the direct, indirect or total effects. Move any mention of structural racism in to the end of the paper when you discuss future research.

*The interpretation in all your other tables flows in a similar fashion. The previous comment on structural racism needs to be removed when discussing tables and spatial correlation needs more explanation. 

*Again, please explain what direct, indirect, and total effects are before discussing all these tables. And, gently remind the reader what that means when interpreting the results.

Discussion:

*Clean up the language a bit. When you state: “The study shows that there is a correlation between deaths and the Hispanic share” , you should write there is a positive correlation between death and the percent of the population that is Hispanic. Or, death increase in PUMAS with a more Latinos.

In the discussion, you suggest that the lack of doctors is a risk factors for foreign born men. This suggest that more doctors are required in the neighborhood. But again, you don’t measure neighborhood. You can say it suggest more doctors in the PUMA or area, but you can’t extrapolate to the neighborhood level. It makes common sense, but the statement goes beyond your data.

Two issues to consider with these risk factors. First separate the risk factors that belong to the PUMA from the risk factors that belong at the individual level. Second, you need citations for some of these risk factors as you list them. Is the research problematic that other scholars have argued that we need to look at these risk factors. As it is written, these risk factors come out of nowhere and are not discussed by their unit of analyses (PUMA vs Individual).

You choose Hispanic Subgroups in NYC that represent at least 4% of the population. Puerto Ricans are the largest Latino group in NYC. Why are you not including Puerto Ricans? This leads me to question the representativeness of your data. If you drop cases or groups from IPUMS, you will need a dummy variable in the analyses which is 1=missing 0=not missing, for instance. If you drop all other groups without this missing variable indicator included in the analyses, you are compromising the sampling integrity of your sample and you cannot make claims about statistical significance. Your p-values will not be valid. So, include that missing value indictor. A more robust way of doing this type of analysis with sampling weights is to use a subpopulation command, where 1 is not missing. The other Hispanic group in your analyses does not represent the remaining 1%, as you state, especially if you exclude Puerto Ricans.

On the Issue of Proxy and Fuzzy Measurement

*Do not use the words neighborhoods or segregation, especially in the abstract. There is a large body of literature measuring what is a neighborhood. The convention is to use either census tracks or census blocks. To suggest that a PUMA is a proxy for a neighborhood is way off base, empirically, conceptually, and theoretically. From a measurement point of view, PUMAS are not neighborhoods. This is especially the case when looked at from the vantage point of construct validity. From a theoretical point of view, equating PUMAS with neighborhoods does a disservice to researchers doing work on neighborhoods and health. Your article will get recognized when people conduct a search for “neighborhoods and health” or related search terms, but it would not be included as part of a literature review on neighborhoods and health. Imagine for a second a meta-analyses on neighborhood effects and COVID. It is unlikely that this study would be included in such an analysis because you are not measuring neighborhoods.

*There is an easy way out of this: Do not use the word neighborhood, especially if your literature review does not review the effect of neighborhoods on health and COVID. What you want to do, instead, is use the word “contextual effects” within a theoretical framework. This way, you can easily make the argument that PUMAS, as macro-level social structure, affect the life chances of individuals. This has the benefit of theoretically grounding your research. I would recommend that you ground your research using the Racialized Place Inequality Framework, highlighted by Velez (2017) and first developed by Burgos, Rivera and Garcia (2017:61 ). I put the references at the end of my comments. Grounding your research theoretically within this framework easily allows you to make the contextual argument about PUMAS. There is another framework that could work by Krieger (2021) called ecological theory, but Burgo’s et.al. framework (2017:61) is more sociologically developed. It also has the advantage of being a framework developed by Latino scholars and we want to promote the work of Latino scholars over the work of their Ivy League peers; that is, if we are concerned about addressing academic inequalities.  

Your paper revolves around identifying risk factors without a theoretical foundation. At the very least, you should discuss how these “risk-factors” fit within a contextual theory, as mentioned above. Regarding risk factor epidemiology, House (2002: 132) argues that “There is an inherent tendency to proliferate an increasingly diverse and scattered set of risk factors, each with modest to small effects, and many lacking a solid evidentiary base as to their impact on health or the degree to which they are distinct from other well-established or putative risk factors for health. An indiscriminately expanding smorgasbord of psychosocial risk factors poses significant problems for the future development of science, practice, and policy regarding the role of social factors in health.”

More recenlty, Keyes and Galea (2017) point out the limits of framing research around risk factors (see references below).

*The way to move out of the risk-factor smorgasbord is to have a small section on theory at the front of the paper and then talk about how the factors you explore fit within that multi-level framework. Only then can PUMA level variables make sense.

Regarding Segregation

There is a large body of literature that pays close attention to the measurement of segregation (e.g., Apparicio et al. 2014). These scholars differentiate between spatial measures of segregation and non-spatial measures of segregation. They also make a distinction between local measures of segregation and global measure of segregation.  In your article, you equate percent minority with segregation, arguing that percent minority is a proxy for segregation.

It is not: percent minority is not a proxy for any measure of segregation.

So, do not use the word segregation to frame your work. Use minority concentration or some other word. The word segregation may get you search hits, but just like the use of the word neighborhood above, you do not measure segregation. As Pinto-Coelho and Zuberi (2015) note, scholars need to be very careful to differentiate between these two terms. For example, if you find that percent minority is not related to your outcome, you cannot argue that segregation does not matter. Why, because you do not measure segregation. So, the terminology greatly matters. I recommend you do not use the word segregation if you don’t measure segregation.

On Structural Racism

You do not measure structural racism and the proxy that you use is not a reliable and valid measure of structural racism. As you know, there is much interest in the study of structural racism and health. That literature shows a positive association whereby higher levels of structural racism are associated with poor health outcomes.

Despite such an association with limited life chances, Republican presidential candidate Tim Scott stated that “America is not a Racist Country” during a recent presidential debate. The color-blind policy implications for racial inequalities are alarming and profound. There is much at stake and properly measuring structural racism is one of the most pressing issues of our time.

Would Tim Scott and his color-blind party be convinced by your measure of structural racism? Would scholars studying health disparities consider your measure of structural racism robust, with high construct validity and reliability? Would race and stratification scholars who are concern about the socio and econometrics of measuring structural racism be convinced by your measures. Unlikely.

*My suggestion is that you drop the word structural racism from the paper since you are not measuring this concept. Using this term loosely minimizes and undermines the fight for racial equality in the health field and the study of racial stratification. Again, move any mention of structural racism to the discussion and argue that we need to explore the relationship between structural racism and COVID.

Here are some articles on how to properly measure structural racism:

Wizentier, M. M., & Stephenson…, B. J. K. (2023). The measurement of racism in health inequities research. Epidemiologic ….

 

Torres Stone, R. A., Ahlgren, N. A., & Bergmann, P. J. (2023). Multiple measures of structural racism as predictors of US county-level COVID-19 cases and deaths. Ethnic and Racial Studies, 46(5), 832-853.

 

Furtado, K., Rao, N., Payton, M., Brown, K., Balu, R., & Dubay, L. (2023). Measuring Structural Racism. urban.org.

 

Dyer, Z. (2022). The Structural Racism Effect Index: A Multi-Dimensional Tool to Measure Neighborhood-Level Structural Racism. repository.escholarship.umassmed ….

Again, you do not measure structural racism so focus on contextual effects. You argue that

“The article shows neighborhood level census data (IPUMS-USA) for New York City to measure spatial segregation and neighborhood characteristics as proxy variables for structural racism.”

*Again, you have no measure of spatial segregation and you do not proxy structural racism. Get rid of this sentence and reframe the discussion around context.

You need more clarity on your problematic. It’s OK to say that scholars have not looked at the risk factors you identify in relation to COVID, a la classical risk factor epidemiology.  Is your research problematic addressing a debate in the literature? Is the problematic that contextual effects matter and people in past have ignored context? As Keyes (2017) notes, it’s often not enough to search for variables to justify an article? Why do the variables that you choose inform a larger problematic? Closely related, why are the variables that you choose theoretically important, and how do they fit within a larger multi-level framework, or any theoretical framework?

References:

Apparicio, P., Martori, J. C., Pearson, L., Fournier, E., & Apparicio, D. (2014). An Open-Source Software for Calculating Indices of Urban Residential Segregation. Social Science and Computer Review, 32(1), 117-128.

 

 

Burgos, G., Rivera, F. I., & Garcia, M. A. (2017). Towards a Theoretical Framework for Assessing Puerto Rican Health Inequalities: The Racialized Place Inequality Framework. Centro Journal: Journal of the Center for Puerto Rican Studies, XXIX(III), 36-73.

 

House, J. S. (2002). Understanding Social Factors and Inequalities in Health: 20th Century Progress and 21st Century Prospects. Journal of Health and Social Behavior, 43(2), 125-142.

 

Keyes, K. M., & Galea, S. (2017). Commentary: The Limits of Risk Factors Revisited: Is It Time for a Causal Architecture Approach. Epidemiology, 28(1), 1-5.

 

Krieger, N. (2021). Ecosocial Theory, Embodied Truths, and the People’s Health. Oxford University Press.

 

Pinto-Coelho, J. M., & Zuberi, T. (2015). Segregated Diversity. Sociology of Race and Ethnicity, 1(4), 475-489.

 

Velez, W. (2017). A New Framework for Understanding Puerto Ricans’ Migration Patterns and Incorporation. Centro Journal, XXIX(III), 126.

 

 

 

 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Author Response

1. Briefly explain what the Hispanic Paradox is..

Answer: Thank you for the comment. This is done between lines 37 and 43.

2. Break the Hispanic paradox to the reader..

Answer: Thank you for the comment. This is done between lines 40 and 43.

3. Be more precise..

Answer: Thank you for the comment. This is done in line 38.

4. Please clarify the unit of analysis.

Answer: Thank you for the comment. This is done in line 40.

5. You should consider abandoning the risk factor framing..

Answer: Thank you for the comment. This is done during the entire paper. The theoretical explanations are summarized between lines 60-97. The empirical explanations are summarized between lines 98 to 107.

6. Hispanic paradox: You mention…

Answer: Thank you for the comment. The Hispanic paradox is now explained as it has been studied in the literature, looking at health outcomes where it applies or not. The importance of PUMAs has been now introduced earlier, lines 84-96 where RPIF is presented, and it is argued PUMAS are a macro level variable. The importance of PUMA’s is also presented from the point of view of spatial correlation (lines 94-97).

7. Structural Racism.

Answer: Thank you for the comment. The references to structural racism and implicit discrimination have been eliminated. The paper now focuses on Segmented Assimilation Theory and RPIF (lines 134-155).

8. Spatial Correlation.

Answer: Thank you for the comment. This section has been expanded to better explain the studies showing correlation at the level of county and how spatial correlation may be related to RPIF (lines 167-189).

9. Heterogeneity

Answer:  Thank you for the comment. In the specific section, heterogeneity in results is presented as part of the health outcomes literature (lines 190-208). Heterogeneity in outcomes is also now discussed between lines 44 to 74, as being potentially explained by Segmented Assimilation Theory, as well as how it is measured in the paper.

10. Materials and Data Sources

Answer: Thank you for the comments. These sections have been modified, explaining the data sources, and presenting preliminary results using maps for deaths at PUMA level, as well as data on spatial distributions at PUMA level for the different Hispanic subgroups and by gender. It also clarifies how the process of cross-walking is done. It introduces now Figures 2, 3 and 4, which are maps with all the information at the PUMA level and with references to the five boroughs of NYC. The section clarifies the number of PUMAS and the software used to manage the data.

11. Results

Answer: Thank you for the comments. The section has been modified, explaining in more detail the results, and giving an interpretation within the context of Segmented Assimilation Theory and RPIF. This is done for each Hispanic subgroup, males, and females.

12. Discussion

Answer: Thank you for the comments. The section has been modified, discussing the results regarding other studies done about COVID 19, and with regard to what Segmented Assimilation and RPIF imply.

 

13. Other comments.

Answer: Thank you for the additional comments.

1. We have dropped the usage of the word neighborhood, and we are now defining clearly what PUMAS are.
2. We are using now an approach based on theory (Segmented Assimilation and RPIF) and not only risk factors.
3. We are not talking of segregation, except when necessary and appropriate.
4. We are not talking about structural racism.
5. We are now focusing in three research questions that explore how the data demonstrates or not the different theories applied. (Segmented Assimilation and RPIF).

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The article would benefit from further condensation and reduction in the number of tables. 

Author Response

1. Thanks for the comment. We have reduced two tables that provide little marginal content to the paper. The number of Tables went from 10 to 8.

Reviewer 2 Report

Comments and Suggestions for Authors

 

The addition of theory at the front end of the paper enhances the manuscript and moves us away from the mind-numbing risk factor epidemiological approach that plagues the social science literature.

 

The authors do a nice job of revising the theory in relations to the findings. They also revisit the theories in the conclusion to bring the manuscript to a theoretical full circle.

 

In the front end of the paper, for the first research question, bring it full circle? Explicitly tell us how the theory informs the first research question. A sentence or two suffices.

 

For the second research question, make it the second question the topic sentence of the paragraph. The tell us how this study informs that second research question. This is just a grammatical fix.

 

Similarly, for research questions 3 and 4, start the paragraph with the question. The rest of the paragraph should justify why the question is important. For instance, it hasn’t been answered, it provides new insights into ?? .. .  and so on. This also make it clear what is the research questions and what is the contribution. Right now, the research questions are embedded into the paragraph and sort of get lost.

 

 

I’m glad to see a Map of PUMA’s and that the authors do not equate PUMAS with neighborhoods. I am also glad to see that the language of segregation has been cleaned up since the authors do not measure segregation. Thus, they fixed the critical theoretical and conceptual issue with the last framing of the paper. The first draft of the paper implied that the authors were looking at neighborhood effects and the effects of segregation. They did not do this in the first version. This second version is thus much improved on this frong.

 

 

In the data and methods section, the authors do a nice job linking the theoretical expectations to the models. It is much improved from the first version.

 

In the discussion, it might be prudent to restate the research questions and provide a broad summary of what the findings show in relations to the research questions. What do we learn? What’s the take home message for each question. Many readers who might be pressed for time will read the abstract and then jump to the conclusion and discussion. So, the conclusion and discussing is a good place to bring the research questions and answers together in a broad sense.

 

Replace the word “bad” with another word. You can use “economically disadvantage” if that is what you mean by “bad”. Be precise.

Comments on the Quality of English Language

Give the paper a grammatical review. For instance, make the research question a topic sentence and explain why it' important. Check spelling throughout. Replace the word "bad" with another word (e.g., economically and socially disadvantaged are).

Author Response

1. In the front end of the paper…

Answer: Thanks for the comment. We have modified the paragraph, and the research question starts the paragraph (see lines 45-70).

2. For the second research question…

Answer: Thanks for the comment. We have modified the paragraph, and the research question starts the third paragraph (see lines 71-77).

3. Similarly for research questions 3 and 4:

Answer: Thanks for the comment. We have modified paragraphs 4 and 5, to start with the research questions (see lines 78-112).

4. In the discussion…

Answer: Thanks for the comment. We have modified the discussion section to provide direct answers to each of the fourth research questions. (see lines 620-676).