DISCLAIMER #1: Given time constraints, the language below is very direct/curt. This is exclusively attributable to wanting to detail the greatest number of proposed edits in the limited time I have allocated to complete this and is NOT indicative of my opinions on the manuscript. I think this is extremely well done and want to make sure that the curtness of my comments are not mistaken for disapproval.
DISCLAIMER #2: Sorry for any grammar/spelling issues
Big picture
My central comment pertaining to your abstract/intro/lit review/theory sections is that all the right pieces are there, but I think you need to be more direct in putting up “sign posts” for the reader to follow your reasoning. You need to be more explicit in laying out your argument and the logic underpinning it. Put simply, you need to use more of an active voice. I also think there are some bits you could trim that are not essential to supporting your argument. On another note, I think you need to name-drop fewer studies and just cite them parenthetically when thematically characterizing existing work. When you do describe particular studies, your descriptions should be shorter, focusing narrowly on the key takeaways that are relevant to your argument. Below, I propose a template for your lit review that maintains an active voice and places “sign posts” for the reader to easily track your reasoning.
Below I will propose a revised theory—and by extension, some revised analyses. The first half of this theory is almost identical to what you have now, so even if you don’t like the proposed changes, much of my comments should still be useful to you.
Turning to your analyses, I think you handicapped yourself by (1) coding your independent variables as categorical vs ordinal, as this dramatically reduces your power in your already small samples, and (2) by analyzing the 2016 and 2020 surveys separately versus pooled. If you continue with the ANES, I think that you should revise these two items.
Outline of new argument
Summary of existing work (no real changes from what you have now)
Immigration is heavily stigmatized.
Latinos are stereotyped as immigrants, which, because of its stigmatization, carries negative social consequences (e.g., xenophobia towards Latinos). Research on the political implications of this stigmatization for Latinos is limited, and even less is known about how this varies by indicators such as pan-ethnic group consciousness, geography, and generational status.
SIT posits that we derive self-esteem from group identification. Further, when our group is attacked, SIT posits that strong- and weak-identifiers react differently. Strong identifiers “lean in” to their attacked characteristic while weak identifiers distance themselves from it. This has a straightforward application to the case of Latinos and social stigma surrounding immigration: strong-identifiers should “lean in” to their connection with immigrants while weak-identifiers should distance themselves from it.
How would we expect this to manifest politically? One obvious way pertains to attitudes on immigration. A second way pertains to support for parties broadly recognized to be pro/anti-immigration. Put simply: strong-identifiers should have more liberal immigration attitudes and be more supportive of the Democratic Party, while weak-identifiers should have the inverse. Indeed, some existing work suggests this is the case (Perez 2015; Garcia-rios et al. 2019).
What Jennifer is adding
One way the relationship between identity strength and political attitudes may vary is by geography: (1) In border states/regions, the policing of immigration is very visible and anti-immigration rhetoric is ubiquitous in local politics. Therefore, stigmatization—and by extension, the political reactions of strong- and weak-identifiers—may be stronger there. (2) In addition, Latinos may also be reminded of this stigmatization in geographic regions with high proportions of foreign-born Latinos. Therefore, stigmatization—and by extension, the political reactions of strong- and weak-identifiers—may be stronger there.
A second way by which the relationship between identity strength and political attitudes may vary is by generational status. Weak identifying immigrants should have sympathy for people like them, and hence their weak pan-ethnic identification should not manifest in conservatism on immigration and republican partisanship. Conversely, later-generation weak-identifying Latinos, more distanced from the immigration experience, should be more likely to let their weak-identification manifest in conservative attitudes/GOP partisanship.
NOTE: I can actually see the “proportion of foreign-born Latinos” argument going the other way. Latinos living in areas with lots of foreign born Latinos may be more sympathetic to them. Maybe you could run your models first, and then post-hoc revise your theory to fit your results :)
Expectations
H1: Pan-ethnic group consciousness is positively associated with more liberal attitudes on the border and support for the Democratic Party among Latinos.
H2: This relationship is strongest in border regions.
H3: This relationship is strongest in geographic areas with high proportions of foreign-born Latinos. [Or the inverse…see earlier note…either way, this is a super interesting piece of geographic variation no matter what you argue]
H4: The positive association between weak-identification and conservative immigration attitudes and Republican Partisanship is driven by later-generation Latinos.
Abstract/Intro
If you go this route, revise your abstract/intro to basically summarize the previous section + talk about methods/findings.
Lit review
NDISCLAIMER #1:: Here is my proposed re-working of the lit review. I think that your first two sections are great, and my edits are mostly geared towards making your argument more explicit. However, I think the remaining sections of your literature review are too long, describing findings that aren’t directly supporting your argument.
DISCLAIMER #2: I kind of ran out of time near the end, but you’ll get the idea.
Stigmatization of immigration
Thee is lots of stigma around immigration. For example, survey reveals that Americans [whites?] associate immigration with negative things like X, Y, and Z. Americans are more likely to stereotype Latinos as immigrants because of factors such as the high proportion of foreign born Americans among this group (cite), the large proportion of US immigrants coming from Latin America (cite), and those subject to immigration enforcement being overwhelmingly Latino (cite). Given the negative stigma associated with immigration, Latinos being broadly stereotyped as immigrants has predictable social implications. Media coverage among Latinos is often negative (cite) and Latinos are often the subject of xenophobic political attacks (cites). While some studies have examined how Latinos respond politically to these attacks, research on this front is limited. Even less is known about these reactions vary by indicators such as pan-ethnic group consciousness and one’s proximity to the immigration experience.
SIT, Latinos, and immigration
According to social identity theory, humans are predisposed to view social group memberships as central to one’s self-conception, and seek to engage in favorable inter-group evaluations to improve one’s self-esteem (cites). How individuals react when their social group is attacked varies systematically by the strength of their attachment to that social group (cite). Individuals with strong group attachments are often emboldened by these attacks, “leaning in” to the attacked characteristic. Among individuals with weak social group attachments, however, the inverse appears true. That is, weak identifiers may distance themselves from the attacked characteristic to avoid psychological discomfort stemming from associating oneself with the content of the political attack.
This model has a straightforward application to social stigma rooted in immigration directed towards Latinos. For Latinos, pan-ethnic group identification is often captured using measures such as [insert examples]. According to SIT, how Latinos react to border stigmatization should vary systematically by identity strength: weak-identifiers should distance themselves from immigrants while high-identifiers should do the inverse.
What are the political implications of these reactions by high- and low-identifiers? One expectation appears obvious: it will impact their political attitudes on immigration. However, this could also manifest in their political partisanship. In recent years the Democratic Party has become more liberal on the issue of immigration. Many Latino voters appear to be aware of this: liberal immigration attitudes appear positively correlated with Democratic partisanship and support for Democratic candidates among this group. Therefore, when faced with stigma about immigration, weak identifiers can be expected to become more hawkish on immigration and more supportive of the Republican Party, and vice versa. Indeed, these expectations are consistent with some existing research. For example, an experiment by Perez (2015) finds that Latinos with weak pan-ethnic group attachments distanced themselves from the group when exposed to xenophobic political attacks. Further, Garcia-Rios et al. 2019) find that views of Donald Trump—whose political tenure was defined by his hostile rhetoric towards Latinos (cites)—vary systematically by identity strength in a similar way, with high-identifiers offering less favorable evaluations. Therefore, identity strength appears to moderate Latinos’ reactions to border stigmatization.
Variation in this relationship (i ran out of time lol)
I argue that the experience of this stigma around immigration varies systematically by three indicators one’s proximity to the immigration experience: (1) generational status, (2) proximity to the border, and (3) the proportion of foreign-born Latinos within a geographic region.
Generational status. Among weak-identifiers, immigrants are less likely to become conservative on immigration because they have sympathy for immigrants.
Geography. Immigrants are heavily demonized in border regions. [expand a bunch on this.] Therefore, the relationship between identity strength and attitudes/PID should be strongest here. In addition. [defend the % immigrant argument you ultimately end up making]
Methods/Results
Dependent variables. Be sure to add pid/vote choice as models. This will leave you with four outcome variables: (1) wall, (2) tighten border security, (3) 7-pt PID, (4) supporting Trump in 2020.
Independent variables. Your sample is too small for categorical variables. Make them ordinal like how they are coded below. (Or honestly, even dichotomous, especially since you may have interaction terms in your models.)
Data. Pool ANES 2016 and 2020. This will give you larger N. You can always put individual models in the appendix.
Coding. Re-code all IVs and DVs between 0 and 1 so the models aren’t buttcheeks to interpret.
BELOW YOU WILL FIND SOME EXAMPLE ANALYSES. NOTE THAT DUE TO LIMITED TIME I DID NOT COMPLETE EVERYTHING ABOVE. ACCORDINGLY, DO NOT PUT ANY STOCK IN THE RESULTS WHATSOEVER AS THEY COULD BE COMPLETELY DIFFERENT. (ALTHOUGH THE GENERATIONAL STATUS RESULTS ARE PROMISING.) INSTEAD, THEY ARE MUST MEANT TO BE AN EXAMPLE OF HOW I’D SET UP THE MODELS RE THE INTERACTION TERMS, AS WELL AS HOW YOU COULD DISPLAY THE RESULTS.
