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

[jialeCharloote]

You used complex system modeling within the framework of urban scaling theory to predict differences in implicit biases between cities. Can you clarify how urban infrastructure, at a city-wide level, ties into individual psychological outcomes such as implicit racial biases? Are there certain urban design features that are more influential?

[PaulaTepkham]

Thank you for sharing your paper with us. The model shown that cities that are more populous, more diverse, and less segregated are less biased, which makes sense to me. However, it seems arguably that because of the size and diversity of the city that drive the level of biases or the city with less biased manner lure more people with diversity in.

[PaulWang-Uchicago]

Thanks for sharing. How might cities with higher levels of racial segregation enhance overall health and income levels, and what challenges could arise in achieving these outcomes?

[graceshaoy]

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?

[mikezheng2001]

During your talk, you mentioned that larger, more diverse, and less segregated cities tend to have lower levels of implicit racial biases. Could you elaborate on the specific mechanisms through which urban structures and social interactions contribute to these observed patterns?

[AnniiiinnA]

Thank you for your insightful paper! The study suggests that city size, diversity, and segregation significantly impact implicit racial biases. How might policies aimed at reducing segregation or increasing diversity in smaller cities be tailored to achieve similar reductions in bias levels as observed in larger, more diverse urban areas?

[ralphluo11]

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?

[Zhang-snow-shirley]

How might the social structure and demographic makeup of cities influence the implicit biases of individuals and groups, and how can urban planners and policymakers use these findings to design more inclusive urban environments?

[hailhan]

I am curious about the rigor of IAT results as a measure of implicit bias. It is probably the pre-eminent source of data available on the topic, but I wonder whether the results are influenced by test-takers' awareness of what the test is measuring? Are there other ways to measure implicit bias so as to reduce reactivity from participants?

[wendixue]

Thank you for sharing this inspiring research! What makes me wonder a bit is whether the choice to use individual averages to represent implicit bias scores at the city level is the best choice. Can the nature of cities with more extreme values and cities with more neutral values be seen as similar?

[fabrice401]

Thanks for sharing this work. While the study suggests rapid learning in response to changing social conditions at short time scales, it also finds evidence for biases influencing social environments at longer time scales. What mechanisms might explain the observed bi-directional relationship between changes in social environments and implicit biases at longer time scales?

[Joycepeiting]

Thanks for sharing. I am wondering: (1) Can the model be extended to other ethnic groups? If so, what differences or similarities do you expect to find out? (2) Interaction patterns may differ for people at different ages. As the sample may be not representative, what kind of metrics will you propose to identify the differences in future analysis?

[AnqiW02]

Thank you for sharing this insightful research! After reading the article. I'm wondering how local cultural factors and history shape the level of implicit bias in different urban areas? And based on this understanding, how can we apply this model to explore the interactions between national social organizations and these biases?

[AnnieZzz01]

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?

[ethanjkoz]

I think this paper proposes a strong argument as to why it is important to look beyond individual level reasons for the persistence and strength of implicit biases. Specifically, we ought to keep in mind the socio-environmental factors like where we live and who we surround ourselves with. This paper deals specifically with racial implicit biases, but in an American context, one cannot talk about race without also bringing up social class. In doing so, it would be interesting to see and compare the results from papers tracking different types of biases and the effect of living in large cities has on them. My question then becomes, when we factor in both class and race as forms of biases, what effects do large cities pose? What type of interaction effect, if any, do these variables have on each other? Does bias against social class have similar patterns to that of implicit racial biases?

[nitomanto]

I was very curious about the measures of segregation that were used in your paper. The paper contains four different measures of residential segregation; I am confused as to why the mean deviance measure would be a useful parameter, as it is immediately followed by the normalized segregation index, a normalized measure that accounts for neighborhoods of dffierent sizes.

[gtriberti9]

Regarding the use of IAT. Does it start showing one of the images and then the other? or does it randomly choose one of the two images to ask within the representative sample? Additionally, what could be the practical implications of your study findings in terms of urban interventions to reduce implicit racial bias in other cities?

[ecg1331]

Thank you for sharing your research!

I have a question about the population of your study and research. It sounds like you were only able to look at 'white' vs 'black' communities based on the census data and IAT study, do you think that this study would produce similar results if you were able to account for all races? And do you see yourself trying this in the future?

[Gio-Choi]

Thanks for sharing. In addition to the regression model, are there any other models you would like to explore? Additionally, are there any specific data types (e.g., geospatial data) you would consider incorporating into the analysis?

[carrieeehuang]

Thank you for sharing this research! Your model emphasizes the role of city-level structural factors such as population size, diversity, and segregation in shaping implicit racial biases. However, cities are dynamic and diverse in ways beyond these variables. How might other factors, such as historical racial legacy, policing, or specific cultural practices, further interact with or moderate the effects of population, diversity, and segregation on implicit biases? Are there specific mechanisms or urban policies that could amplify the positive effects you observed, and how would you prioritize these factors for future research?

[yangyuwang]

Thanks for sharing the research on how social contexts in the city-level would influence the implicit biases!

My questions would be regarding the measurement of Implicit Association Test (IAT). I have read several papers utilizing IAT to measure some kinds of implicit attitudes, such as towards cooperation (Srivastava & Banaji 2011). But, I wonder how valid and reliable IAT is and to what extent it could be leveraged in other types of research.

For the former question, as IAT utilizes the differences in response time towards different association of words to measure the bias, it could still be influenced by certain social desirability and might not only result from the two words shown on the screen, but also links towards other words (like "white" might be linked to "white things", instead of a race). And most important thing is that IAT could have large variation even in one individual across different tests. That is why I doubt the reliability of the measurement.

The latter question is in a more positive way of thinking. Can IAT be used in other situations, like attitudes towards LGBTQ+ groups or environmental issues? If you have read some papers about that, please share them with me.

[ymuhannah]

Thanks for your sharing. Your model highlights the relationship between city characteristics and implicit biases, with larger, more diverse, and less segregated cities showing lower biases. Could you elaborate on how this model might account for the role of digital social networks, which can transcend geographical boundaries? Do these networks amplify or mitigate the effects of city-based social interactions on implicit biases?

[deyuz003]

Thank you for sharing!

You mentioned in the paper that population, diversity, and segregation are three structural factors in the model. In your opinion, what other significant factors might be potentially included in the model?

[robertorg]

Thank you for sharing your research! I have two questions:

First, in your paper, you use racial categories like ‘Black’ and ‘White,’ which are well-defined in the U.S. context. However, in societies where racial identities are more fluid and can shift depending on social, cultural, or situational factors, such as language, socioeconomic status, or region, a person might be seen as ‘white’ by some and ‘indigenous’ by others. Do you expect that the mechanisms driving bias and how to decrease would still apply in these contexts?

Additionally, is it possible that living in more diverse cities influences not the actual level of implicit bias, but rather how individuals respond to the IAT due to increased awareness or more frequent inter-group interactions?

[Mengyang-Cao]

Thanks you for sharing your research! A quick question: How do you plan to investigate whether and how cities systematically facilitate interactions of differing quality

[Lezy01]

Thank you for sharing this paper! It was both interesting and enlightening to read an article in a field I haven't been exposed to before, especially as it addresses significant social phenomena. Since I have no prior experience with theories in psychological and neural processing, I found it challenging to fully grasp the theoretical model used to illustrate the question.

Firstly, I’m curious about the definitions of the between-group and within-group relative rates of interaction. Why are these parameters important when calculating the number of per-capita interactions? Secondly, I’d like to understand why the diversity and segregation adjustments take the specific forms mentioned in the article. Is there an intuitive explanation for this model setup?

[bkehoechicago]

Thank you for sharing your research.

I have some concerns about the validity of the Implicit Association test (see here and here). How would you go about addressing these concerns in relation to your research? Are there alternative measures you could use?

[ZheruiLU]

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?

[HamsterradYC]

Thanks for your sharing this. I'm wondering daily social interactions within urban environments can inadvertently reinforce or mitigate implicit biases, closely linked to the racial and cultural diversity of the city. I am very interested in this and would like to know whether this pattern can be effectively extrapolated to urban environments across different countries and cultures. Are there international comparative data supporting this theory?

[harringtonfan]

Thank you for sharing! I'm particularly interested in the section named "Timescales of temporal precedence". I think one of the puzzles you grapple with here is essentially a chicken-egg question: is the structural change of the urban settings preceding the change of implicit bias or the other way around? And according to your findings, it highly depends on the span of timescale: mutually equal direction take place more on longer timescale, whereas the structural change preceding the implicit bias more often take place at 1-2yr time scales. My question (or a rough thought, so to speak) is, despite that the racial bias is more or less ubiquitous around the world, especially being a concern in the multiracial societies, the variation of the state or the governance could play a big role in that. For example, Singaporean government introduced the mandate named Ethnic Integration Programme in 1989 to enforce ethnic makeup of each residential unit in order to facilitate the inter-racial interaction so as to eliminate the bias over time. So I think, given the governmental intervention, the case of Singapore may not fit well in this finding, which is built upon the individuals' spontaneous behavior. But I just wonder if there's any specific policy made by any state government in the U.S. (e.g. imagine, CA. Governor Newson issues a certain housing policy to dilute the density of the population in Chinatown) that somewhat resembles Singaporean intervention that may leave a profound legacy, if so, how can we make sure it's considered in the research and maybe "controlled"?

[VeraMao]

Thanks for sharing! These findings really got me thinking about the nature vs. nurture debate. I am curious about how might the relationship between city size and implicit bias challenge our assumptions about urban versus rural attitudes towards race? And what ethical considerations arise when considering urban planning as a tool to influence implicit biases?

[zhian21]

Thanks for sharing! The paper explores how urban characteristics affect implicit racial biases, revealing that larger, more diverse, and less segregated U.S. cities tend to have lower bias levels. Then, by using urban scaling theory and analyzing Implicit Association Test data, the study also demonstrates that citywide social contexts play a significant role in shaping individual biases. Given these findings, how might targeted urban policies that promote diversity and reduce segregation further accelerate the reduction of implicit biases in cities over time?

[QIXIN-LIN]

Thanks for sharing! I'm wondering how the policy would affect that. Also, is there any policy that can lower implicit racial biases based on this paper's findings?

[natashacarpcast]

Thanks for sharing. Regarding the third hypothesis: "that less segregated cities have lower levels of implicit biases" , I'm curious of why wasn't this a given in the first place? As far as I understand, segregation is often a consequence of racism/bias/discrimination, so I wouldn't have doubt them being correlated. Can you explain why this was an hypothesis, instead of a fact?

[HzSeaski]

Thank you for sharing this great research! I’m very curious about the role that contemporary geo-political factors that influences social context to some extent. Findings suggest that social policies aiming to produce diversity and reduce segregation can potentially change social context and reduce implicit racial bias. My question is, at a cross-city level, would you predict that a more diverse, less segregated city with a larger population might not only influence residents' cognition and perceptions of their own social context, but also extend these effects to their views of neighboring cities? Could this lead to the development of new implicit biases, or the reinforcement of existing ones, toward individuals from adjacent cities, especially if those cities differ significantly in their levels of diversity and segregation?

[hchen0628]

Thank you for sharing. If larger, more diverse, and less segregated cities can reduce implicit racial biases, can we systematically reduce these biases through conscious urban planning and policy interventions? What challenges might arise when implementing such strategies?

[kexinz330]

Thanks for sharing this excellent study! You mentioned that the current model does not directly address the quality of interactions between groups. Considering that interaction quality could have a significant impact on bias, is there any future research planned to explore how different types of interactions (e.g., in workplaces, communities) might influence implicit bias?

[MaoYingrong]

Thank you for sharing your work! My questions are: With increasing urbanization and technology use, such as social media, how do you see the relationship between urban scaling and implicit biases evolving? Do virtual interactions impact implicit biases in similar ways to physical urban environments?

[CallinDai]

Thank you for sharing! Since the paper focuses on aggregate-level patterns of inter-group interactions across entire cities, do you think it is possible that in more diverse, less segregated cities, these interactions predominantly occur among specific subpopulations, as mentioned, younger, more educated, and predominantly female individuals, which are overrepresented in the IAT sample? If so, could this concentrated interaction pattern introduce bias into the model’s results and limit the generalizability of the findings to the broader population, potentially threatening the validity of the conclusions?

[gabereichman]

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?

[CallinDai]

'm wondering how local cultural factors and history shape the level of implicit bias in different urban areas? And based on this understanding, how can we apply this model to explore the interactions between national social organizations and these biases?

Hi, I think you raised some very good points.

Thank you for sharing your research! I have two questions:

First, in your paper, you use racial categories like ‘Black’ and ‘White,’ which are well-defined in the U.S. context. However, in societies where racial identities are more fluid and can shift depending on social, cultural, or situational factors, such as language, socioeconomic status, or region, a person might be seen as ‘white’ by some and ‘indigenous’ by others. Do you expect that the mechanisms driving bias and how to decrease would still apply in these contexts?

Additionally, is it possible that living in more diverse cities influences not the actual level of implicit bias, but rather how individuals respond to the IAT due to increased awareness or more frequent inter-group interactions?

Hi, I think you raised some very good points here. Especially given that more diverse and less segregated cities are possibly also more economically developed or educated, people’s perception of racial categories might be shaped not just by physical appearance but by other factors, the mutual reasoning here is to be sure to take into consideration.

[Jessieliao2001]

Thanks for sharing, my question is: In cities where implicit biases are lower due to population size and diversity, have you found any correlation between reduced biases and economic growth or productivity? Does a reduction in biases contribute to better economic outcomes, and if so, through which mechanisms?

[Weiranz926]

Thanks for sharing！What are some potential long-term social or economic impacts you foresee if cities with high segregation levels successfully reduce their segregation in line with your model's predictions?

[Daniela-miaut]

Thank you for the inspiring research! I really like the idea to use implicit biases tests. I'm just curious if you are thinking of doing such research on a micro level, say, on the local community level. Would the results be different if the between-group interactions become deeper in daily lives (not just encountering in the same city)?

[anzhichen1999]

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?

[zhuoqingli526]

Thank you for sharing your work! Do you think the model you've developed could be adapted to explore implicit biases beyond race, such as biases related to gender or socioeconomic status, within urban environments?

[jstanleyi]

The workload of this paper is tough to imagine as it contains so much compounded analysis. Although as a mathematical background student, it seems that I still need spend lots of efforts in understanding the formulation. Therefore, I am curious about the process of how such complicated formula was constructed. Meanwhile, I find it interesting that this paper use Noise Ceiling as a measure. I am curious about its special advantage other than some normal checks in significance.

[Daniela-miaut]

Thanks for sharing! The paper explores how urban characteristics affect implicit racial biases, revealing that larger, more diverse, and less segregated U.S. cities tend to have lower bias levels. Then, by using urban scaling theory and analyzing Implicit Association Test data, the study also demonstrates that citywide social contexts play a significant role in shaping individual biases. Given these findings, how might targeted urban policies that promote diversity and reduce segregation further accelerate the reduction of implicit biases in cities over time?

I am also curious about that! Also, thinking of persuading the policy makers, Prof. Berman, how would you quantify this effect in plain words? Like, in a way that we can say to convince lay people.

[ubaidjamal0]

The idea that larger, more diverse, and less segregated cities have lower levels of implicit bias opens up new avenues for urban policies and interventions. My question, then, is how can one account for resistance by certain groups towards policies and interventions that encourage intergroup interactions?

[nourabdelbaki]

Hi Prof. Berman, Thank you for sharing your paper with us. I am very excited to hear your talk! 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?