Questions for Marc Berman on his 10/10 talk on "Implicit racial biases are lower in more populous more diverse and less segregated US cities"

[muhua-h]

Pose your questions as Issue Comments (below) for Marc Berman regarding his 10/10 talk on "Implicit racial biases are lower in more populous more diverse and less segregated US cities".

Abstract - Implicit biases - differential attitudes towards members of distinct groups - are pervasive in human societies and create inequities across many aspects of life. Recent research has revealed that implicit biases are generally driven by social contexts, but not whether they are systematically influenced by the ways that humans self-organize in cities. We leverage complex system modeling in the framework of urban scaling theory to predict differences in these biases between cities. Our model links spatial scales from city-wide infrastructure to individual psychology to predict that cities that are more populous, more diverse, and less segregated are less biased. We find empirical support for these predictions in U.S. cities with Implicit Association Test data spanning a decade from 2.7 million individuals and U.S. Census demographic data. Additionally, we find that changes in cities’ social environments precede changes in implicit biases at short time-scales, but this relationship is bi- directional at longer time-scales. We conclude that the social organization of cities may influence the strength of these biases.

Reading List:

• Stier, A.J., Sajjadi, S., Karimi, F. et al. Implicit racial biases are lower in more populous more diverse and less segregated US cities. Nat Commun 15, 961 (2024). https://doi.org/10.1038/s41467-024-45013-8

[MaxwelllzZ]

Hi, Prof. Berman, I am curious about a question related to Measurement Reliability: You mentioned that implicit biases are inherently noisy attitudes, which impacts the R² value even in a well-fitting model. Could you elaborate on the specific challenges this noise presents when trying to validate causal interpretations across different urban contexts? What steps, beyond noise ceiling estimation, did you take to mitigate this measurement issue?

[yifand1023]

Thank you for sharing this article, I saw that the implicit biases in this article are basically constructed as racial biases. So is it possible to generalize the conclusions drawn and apply them to non-racial factors such as gender, sexual orientation? Would the diversity of the city, population size, and level of segregation correlate with non-racial factors? I think this is a very interesting exploration because race, gender, and sexual orientation are usually explored together.

[ruining-he]

Thanks for sharing!

Since different cities may influence people's implicit bias differently, maybe migration among different cities will also help people decrease their bias (Because people can experience different lifestyles)?

However, people who can migrate between cities may have higher incomes, which may confuse the cause.

[haonan14]

Thanks for sharing! How can increasing diverse interactions in online communities, such as social media fan groups, reduce implicit biases among users over time, and what might be the long-term effects of such changes on community cohesion and support?

[tyeddie]

The method of this paper is novel to me, particularly the way you model human psychology with mathematical functions. In terms of the Implicit Association Test (IAT) used in the study to quantify the level of implicit bias, I am curious how is bias evaluated and what are the parameters involved?

[Adrianne-Li]

Thank you for sharing this interesting paper! Given the findings that implicit racial biases are lower in more populous, diverse, and less segregated cities, what are the potential policy implications for urban planning and development? Specifically, how can city officials leverage these insights to reduce implicit biases in smaller or more segregated cities? Additionally, how does the bi-directional relationship between social environments and implicit biases evolve over longer time scales?

[Cosmo280]

Thank you for sharing!!!

I'm a little curious that what specific mechanisms or interventions can most effectively reduce implicit racial biases in smaller, less diverse, and more segregated cities, beyond increasing population size, diversity, or reducing segregation?

[yilinx-10]

Thank you for sharing this amazing research! When designing and implementing urban policies, what are the impacts of people being aware of policies' intentions to encourage inter-group contacts and reduce segregation? What are the ethical concerns?

I am also curious about the racial IAT data. The respondents were responding to White or Black faces and Good or Bad words. How does one ensure the perception and identification of faces are uniform across respondents? For example, one may recognize a White face as of a specific European descent and have an implicit bias towards that. How does this study take into account situations like this?

Considering societies with a high flow of internal migrations seeking job opportunities in more economically developed regions and low racial tension, what are the takeaways we can get from this study? What alternative measures can be taken to study implicit bias for local v. migrant workers?

[TheBestCoder-1]

Thank you for sharing this wonderful research. Your findings suggest that changes in social environments precede changes in implicit biases at short timescales. Could you elaborate on the mechanisms that might account for this temporal precedence? What specific social changes do you think most significantly influence implicit bias?

[Vindmn1234]

Thank you for sharing the progress on such interesting research area. My question is: How can urban planners practically leverage your findings to reduce implicit biases through city design, particularly in terms of fostering diversity and reducing segregation?

[yuetianyuan]

Does the level of residential racial segregation in U.S. cities prime participants' implicit racial biases, as measured by IAT scores, particularly influencing faster response times associated with in-group (White) versus out-group (Black) pairings?"

[tyeddie]

Given the study's findings that more populous, diverse, and less segregated cities have lower implicit biases, what interventions could urban planners and policymakers design to promote intergroup contact and reduce segregation in cities?

As a non-psychology student, my biggest takeaway from this paper is that implicit biases cannot simply attribute to individuals but it's significantly influenced by the characteristics of the cities. As such, efforts to promote racial equity should not be limited to education but must also include concrete policies that promote intergroup interactions. For example, avoiding racial segregation of any kind, creating more common spaces for diverse communities, and implementing affirmative policies that support immigrants and multiculturalism are crucial steps in fostering more inclusive, equitable environment.

[amritapathak1]

Building on this study's methodology of linking spatial scales and implicit bias through urban scaling theory, how might we extend the analysis to examine the relationship between economic stratification and implicit bias persistence? Specifically, could the model be adapted to test the hypothesis that economic privilege enables selective exposure to diversity, potentially creating differential rates of bias reduction across socioeconomic groups? This could examine whether:

1. Middle-class populations, due to their increased likelihood of workplace and institutional diversity exposure, show different patterns of implicit bias change compared to higher-income groups
2. The relationship between city-level diversity and bias reduction varies across economic strata, particularly in cases where wealthy enclaves exist within otherwise diverse cities
3. The bi-directional relationship between social environments and implicit biases (noted in the time-scale analysis) manifests differently across socioeconomic groups

This extension would help understand how economic segregation might moderate the relationship between urban characteristics and implicit bias reduction, potentially revealing whether the benefits of diverse, populous cities for bias reduction are equitably distributed across economic classes.

[yuhanwang7]

Thank you for this insightful paper. I found the model linking city size and diversity to implicit bias reduction particularly interesting. I have a couple of questions. First, does the use of Implicit Association Test (IAT) data, which skews towards younger and more educated participants, limit the generalizability of the findings on implicit racial biases in cities? Second, considering the model's predictions, do you think urban planners could practically apply this framework to design cities that foster more equitable social environments?

[Charlottefox]

Thanks for sharing! I want to ask about the relationship between bias and city structure. Is bias one of the factors contributing to the composition of city structure? Could people change their place of residence due to bias, thereby reinforcing and dividing the city structure? Thank you

[BaileyMeche]

Thank you Prof. Berman for presenting this work. I see the paper says that more contact with people from different racial groups can reduce bias, but what about when that contact is just surface-level or happens only a few times? In real life, we know biases often come back if people don’t have deeper or ongoing relationships across different groups. How does the model handle the idea that just being around diverse people might not be enough to create lasting change in racial attitudes? Couldn’t modeling the 'right' kind of interactions, depending on the setting (like schools or workplaces), show a more meaningful impact on reducing bias?

[0theinfinite]

Thanks for sharing your paper! My question is about how to caculate individual psychology in this research? How to determine which psychology index contributes to the question and how to get these data?

[ksheng-UChicago]

Thanks for sharing. The incorporation of urban scaling theory in this paper is extremely interesting. What do you think of gated communities in cities? For example, there are increasing numbers of super high-rise residential towers in Manhattan, and Downtown Chicago. Geographic segregation might not be obvious in such a city, but vertical segregation seems more extreme.

[yuy123337]

Hi Professor Berman, as someone new to mathematical modeling, I'm particularly interested in understanding the logic behind designing models that aim to capture implicit bias. From your work, it's clear that you connect inter-group interactions in urban environments with implicit racial biases. I am wondering what is the reasoning behind choosing certain variables (like city size, diversity, and segregation) to link social interactions to measurable biases? Additionally, given that implicit biases are influenced by a wide range of social and psychological factors, how does the model account for potential discrepancies between people's internal attitudes and the biases detectable through tools like the Implicit Association Test (IAT)?

[Brian-W00]

This study has uncovered how the way cities are socially organized influences people’s implicit biases. Then how can we use this insight to design more inclusive and diverse urban environments?

[bkehoechicago]

My question is how might the findings that larger, more diverse, and less segregated cities exhibit lower levels of implicit racial bias inform urban planning and public policy strategies aimed at reducing systemic racial disparities across various sectors, such as education, healthcare, and law enforcement?

From a public policy perspective, if we accept the causal relationship between population density and implicit racial bias, eliminating policies which favor less dense suburbs would work to reduce the level of implicit racial bias.

[Zhuojun1]

Thanks for sharing! Your research highlights the role of complex system modeling in understanding implicit biases within urban settings. How might this approach be applied to other forms of social biases or inequalities? Are there particular aspects of the urban environment that you believe warrant further investigation through this lens?

[zixuanzzx01]

Thank you for sharing! I find it very interesting to conceive of the city as "as social networks enabled and structured by cities’ hierarchical infrastructure networks" (urban scaling theory), which is essentially formulated based on the principle of optimizing spatial cost. I wonder whether there are alternative ways to model social networks within the urban setting? (Should we take the scaling theory as a given?) What about segregation in other spatial settings?

[binyu0419]

Thank you for sharing. I am also using the same ACS dataset, and I am wondering why you are using five-year population estimates instead of one-year estimates.

[ZheruiLU]

I'm a little curious that what specific mechanisms or interventions can most effectively reduce implicit racial biases in smaller, less diverse, and more segregated cities, beyond increasing population size, diversity, or reducing segregation?

In smaller, less diverse, and more segregated cities, I believe that there are several specific mechanisms can be employed to reduce implicit racial biases without directly increasing population size or diversity. One effective approach is to facilitate structured inter-group contact through community programs, schools, or workplaces that encourage meaningful cooperation and interaction between different groups. These opportunities for interaction can reduce biases by fostering familiarity and understanding.

Additionally, implementing evidence-based implicit bias training can help individuals recognize and mitigate their unconscious biases. Such training is particularly effective when reinforced over time, ensuring that its effects are sustained. Media representation also plays a crucial role—promoting positive portrayals of minority groups in local media can create indirect exposure, influencing public attitudes even in the absence of direct inter-group contact.

[ruining-he]

Since the paper finds that larger, more diverse cities have less implicit bias, I wonder what this paper could say about "online" cities -- how might these findings reflect or change for online communities?

I'm also interested in this question! But I think except for professional websites like LinkedIn, other online communities may not have distinct "city facilities" or "infrastructure". Reddit might be a middle one because it is divided into subreddits. So I guess there is some social media research focusing on these issues.

It's also good to hear the opinion of today's speaker!

[gabereichman]

Thanks for sharing this useful information. Implicit Association Test (IAT) can have some measurement inconsistencies, so how do you go about dealing with noise or inconsistencies in the data from the IAT based on the experimental results you give? Are there alternative measures that provide more stable estimates of implicit bias across cities?

This is a great point to bring up with any online reaction-time data that is collected. I would ask even further if there are potential unexpected patterns that could be influential to testing conditions. One example that comes to mind is economic inequality, which could result in systematically different qualities of computers or even levels of distractedness when completing the task that could influence the results and lower the validity of the measure. Population density is another such factor which could influence the level of isolation a participants have when taking the IAT that could also act as a predictor in the model.

[bucketteOfIvy]

This paper focuses heavily on how one aspect of the urban environment---segregated communities---shapes implicit bias. However, segregation often comes with exploitation and differences in quality of life that are visible in the built environments in which communities reside. To what extent do you believe these built environment factors may capture the remaining implicit bias, and to what extent is higher (spatial) resolution data available to test models utilizing these factors?

[yangyuwang]

How could LLMs be employed to predict the long-term evolution of implicit racial biases in urban populations based on historical and demographic data, such as the census data used in this study?

It's a quite good point to utilize LLM to the study. I think simulating the agents by setting city-level structures could be a desirable way to do so. Upon that, we could look at how LLM respond to different settings of popularity and diversity, and test how they interact with racial biases.

[saniazeb8]

Thank you for sharing such interesting work. I think this framework is essential for studying mega cities and opportunities they create while keeping the inequality in mind. I want to learn more on the methods you used. looking forward to it!

[yunfeiavawang]

Thanks for sharing! Your paper models how city characteristics like population, diversity, and segregation levels influence implicit biases. Given your findings that diversity and less segregation predict lower bias, how might temporary increases in diversity, such as through tourism or short-term events, impact implicit bias levels in the short term? Would these temporary shifts produce measurable changes, or are long-term residential patterns more critical?

[JonathanPMonroe]

I appreciate your paper's insights. I am curious about the potential generalizability of your findings. Considering that diverse heterogeneous populations are becoming more common in the 21st Century, will your results apply to other populations beyond the US? Suppose an immigration-positive country has a diverse populous, but a majority culture dominates the political landscape. Will this cultural prevalence in the country's regulative decisions diminish over time as the minority populations grow in size?

[Kevin2330]

Thanks for sharing! Your research suggests that more diverse and less segregated cities show lower levels of implicit racial bias. How do you account for potential confounding variables, such as economic inequality or education levels, which might also influence bias levels alongside diversity and segregation?

[moeee08-cyber]

Your finding suggests that implicit racial biases are lower in more populous, diverse, and less segregated U.S. cities. This conclusion is based on the urban scaling theory at the city level. Do you think this relationship pattern would change when the research unit shifts to a larger scale, such as a country, or a smaller scale, such as regions within a city? What differences might arise, and why?

[Yunrui11]

Your research suggests that city diversity and reduced segregation are associated with lower implicit biases. Could you elaborate on whether specific urban policies, such as housing or zoning reforms, could be directly used to mitigate these biases further in less diverse or more segregated cities?

[ruoting-Y]

How do differences in interaction quality between in-group and out-group individuals in diverse urban areas influence the rate at which implicit racial biases are reduced, and could higher levels of negative inter-group interactions offset the positive effects of city size and diversity on bias reduction?

[Yunrui11]

Thank you for sharing this interesting paper! Given the findings that implicit racial biases are lower in more populous, diverse, and less segregated cities, what are the potential policy implications for urban planning and development? Specifically, how can city officials leverage these insights to reduce implicit biases in smaller or more segregated cities? Additionally, how does the bi-directional relationship between social environments and implicit biases evolve over longer time scales?

Thank you for your thoughtful response! I agree that urban planning could indeed play a key role in addressing implicit biases through policies like inclusive zoning or affordable housing, particularly in more segregated or smaller cities. Encouraging diverse social interactions seems crucial for long-term bias reduction. I also appreciate your point on the bi-directional relationship between biases and environments over time. It’s exciting to think about how these dynamics could evolve and reinforce positive change.

[Adrianne-Li]

Your research suggests that city diversity and reduced segregation are associated with lower implicit biases. Could you elaborate on whether specific urban policies, such as housing or zoning reforms, could be directly used to mitigate these biases further in less diverse or more segregated cities?

Hi Rui, this is an interesting question. As far as I am concerned, urban policies like housing and zoning reforms can help reduce implicit biases by promoting diversity and integration. Inclusionary zoning policies, which create more affordable housing in mixed neighborhoods, can increase contact between different racial and socioeconomic groups, reducing stereotypes. Similarly, mixed-use zoning fosters diverse social environments by integrating residential and commercial spaces, encouraging cross-group interactions.

Creating more shared public spaces can also facilitate positive interactions, further lowering biases. Over time, these policies can gradually shift social norms and reduce implicit biases in less diverse or segregated cities.

[PENGRUISU]

This paper is extremely meaningful and marvelously. Before reading and watching this presentation, I did not expect that this kind of implicit biases can be monitor or test in such ways. This topic is important for the present society. What I expect more may be some individual level investigation. IN ALL, this is a quiet innovative paper and a milestone.

[zimoma0819]

Thank you for sharing the paper. You have mentioned that living in the bigger cities is negatively related to having depression, and you have also stated that for this study you ignored the weight (strength) of relationships between individuals with their social connections. I’m wondering how the results might change if you do consider the strength. People in big cities might have more social connections, but they are not too close to most of them. And maybe close relationships could have more impact on the likelihood of depression.

[PENGRUISU]

This paper is extremely meaningful and marvelously. Before reading and watching this presentation, I did not expect that this kind of implicit biases can be monitor or test in such ways. This topic is important for the present society. What I expect more may be some individual level investigation. IN ALL, this is a quiet innovative paper and a milestone.

I mean there may be a fallacy of composition. Do the author consider about it? Like it seems we reached a conclusion in the society level, but it may not be true for a single person?

[zcyou018]

Thank you for sharing! How could the findings about the relationship between city size, diversity, and implicit racial biases be applied to inform urban policy strategies aimed at reducing biases in smaller or more segregated cities?

[schen115]

Thank you for sharing this impressive research! I was really interested in your findings, particularly how diversity seems to have a bigger impact than segregation in reducing implicit biases in U.S. cities. I was wondering if you could explain how this aligns with psychological theories like Allport’s Contact Hypothesis, especially when it comes to the learning rates you’ve estimated for bias reduction? Additionally, since your research focuses on city-wide factors, do you think AI simulations that model individual-level interactions in diverse environments could provide deeper insights into how biases change at a more detailed, personal level?

[Pritam0705]

Thanks for sharing. Your study finds correlations between city characteristics and implicit biases. How confident are you in attributing causality, and what additional research might help establish causal relationships?

[XiaotongCui]

Could you explain more about the bi-directional relationship you found between city social environments and implicit biases at longer time scales? Are there particular social or demographic trends that you think are more likely to influence these biases over the long term?

[volt-1]

Thanks for sharing! I'm curious about the potential reversibility of these biases. Given that you found evidence of bidirectional relationships between city social environments and implicit biases.

[66Alexa]

Thanks for your insightful research! Your findings suggest that urban environments influence the level of implicit bias. How might urban planners or policymakers leverage this knowledge to design cities that actively reduce racial biases? Are there specific urban design elements that might be most impactful?

[Huiyu1999]

Thanks for sharing! While your results show correlation, are there plans for future research to directly test the causal mechanisms between urban diversity and implicit racial bias reduction?

[yilinx-10]

Thanks for your sharing. The paper shows short-term changes in urban environments impact implicit biases, while long-term effects may be bidirectional. Could this imply that short-term interventions, like diversity training, are more effective than long-term structural changes? How should policies be designed to address these different time scales of influence?

After today's workshop, I think the bidirectional long-term effects mean that the change (reduction) in implicit bias may lead to a social environment that allows more interactions whilst in the short term we only observe the effect of social environment on people's implicit biases.

[BaileyMeche]

There is a brief discussion of the role that the quality of inter-group interactions can play, in particular in relation to the cognitive costs of some of these interactions. I can imagine a scenario in which there is some optimal level of interaction that lowers implicit bias before individuals get too cognitively taxed resulting in diminishing returns or even negative results, as the discussion in the paper suggests. Could it be the case that there is some ideal level of interaction in certain scenarios, and could this model be used or adjusted to accommodate for the potential of a non-monotonic relationship between implicit bias and inter-group interaction frequency?

The paper does touch on the cognitive costs of inter-group interactions, but it doesn't fully explore how the frequency of those interactions might lead to diminishing returns or even negative effects. From what we know in social psychology, too much inter-group contact, especially when it involves high levels of cognitive strain or negative experiences, can lead to burnout or reinforce biases. It’s definitely possible that an optimal level of interaction exists - enough to reduce bias but not so much that it causes cognitive overload. Adjusting the model to account for this non-monotonic relationship could make it more realistic by incorporating a threshold where more contact stops being beneficial and starts having negative effects.