International Policy Center Home Page

Bicentennial Symposium day 2 session 6

November 10, 2017 1:37:23
Kaltura Video

The University of Michigan has long been a leader in social science research on the many dimensions of social inequality. This bicentennial symposium will highlight these contributions by focusing on the work of distinguished social scientists who were trained at the University of Michigan.


0:00:01: Thank you all for being here for the second day of our Bicentennial Symposium: Impact on Inequality. We had a great day yesterday. We got three more excellent sessions today. And it's an honor to introduce here, to make some opening comments, Jack Hu, who's the Vice President of Research at the University of Michigan. He's also Professor of Mechanical Engineering, Industrial and Operations Engineering, and the J. Reid and Polly Anderson Professor of Manufacturing in the College of Engineering. As Vice President for Research, Jack has many responsibilities overseeing the research program of the entire university. And in the Institute of Social Research, we deal with Jack a lot. He's always very supportive, he's held a number positions in the School of Engineering before moving into the Office of Research. His teaching and research interests include manufacturing systems. He does all kinds of very interesting work, has received many awards, and was elected to the National Academy of Engineering in 2015. So, Jack, thank you for coming over this morning.


0:01:21: Good morning.

0:01:22: Good morning.

0:01:23: Yes, nice to see all of you in this cold November morning. Yeah. To all of you are coming from out of town, a warm welcome. Welcome to the University of Michigan. So you said, a great pleasure for me to provide a short set of remarks to welcome you to this bicentennial symposium on social sciences and inequality. You know, the University of Michigan is known for its broad excellence. We have many different disciplines that are all ranked in the top by US News & World Report, as well as any other measures. Our programs in the social sciences, including economics, policy, political science, education, social work, sociology, women's study, and so on, they're all very, very strong. Also our culture of interdisciplinary collaboration, I will say it's organic. Faculty work together without really the need from deans or Vice President for Research or provost and the president to promote that type of work.

0:02:36: But the strength, I think, across the disciplines is a great foundation for us to leverage, to build something bigger, to try to make the sum bigger than the individual adding their efforts together. So the responsibilities of the Vice President for Research actually are in, I will say, three groups of activities. A first one is to support and continue to build interdisciplinary research initiatives that are larger, bigger, and truly Michigan. Michigan initiatives that leverage the broad excellence of research. The second area is to support research, whether it's individuals, small groups, or large activities. We support all research. Of course, across the different disciplines, research type vary. We have health sciences, five, six different schools, engineering, physical sciences, social sciences, humanities, and art. We support all research and scholarship. The third area is to protect university research through our efforts in research ethics and compliance. We have a large team of people who are in this area. Efforts include IRBs, human research protection, animal protection, export control, conflict of interest, responsible conduct of research, many things that are really set up to protect our research efforts.

0:04:08: You may already know that we have a number of institute that are part of the Office of Research, that are very much in the social sciences. Center of Human Growth and Development: We look at human development, impact of environment, impact of family, and of course health-related research effort associated with children's development. Institute for Research on Women and Gender, IRWG. We study women and gender issues. Of course, we have many programs that facilitate faculty collaboration. Last week on Wednesday, we had our annual MCubed symposium. MCubed is an innovative seed funding mechanism that we started in 2012. Faculty members, typically three faculty members from at least two different units, they're given a nominal token of some money, and typically that is $20,000 per token. Of course, last time, a year ago, we started a small token as $5,000. Really that was to enable faculty in the humanities, social sciences, and the arts to collaborate together. So you have a nominal token, but if three faculty members from at least two different disciplines come together and they all agree on a topic of mutual interest. And then you cash your token, your colleagues cash their tokens, then you have three faculty members come together on the same topic, then you have a project. So proposals are really by faculty putting their own money, a portion of that token come from the faculty themselves, from the deans, and also the provost.

0:06:00: So put some skin in the project, in the game, and then the proposals were not reviewed, so automatically the faculty get support. We have programs as such where we enable faculty to collaborate as small teams. Of course, the topic of the symposium, social inequality, is a excellent area for interdisciplinary research and collaboration. Of course, it's not my research, so I can only browse through the topics and the speakers. It seems to me we had outstanding speakers yesterday, and continue today. This is also a reflection, I would say, the Michigan leadership in this area. Not only our faculty who have done outstanding research are speaking, but we also brought many, I think 30 total, former Michigan students, graduate of the university, who went on to become scholars in their own, who have made a impact, and then we invited them back to the university to this symposium. It's also great to see the link between research and policy. In fact, now the university is strongly promoting faculty engagement at national level, state level, and community levels. President Schlissel, during his annual leadership breakfast, so that's about three weeks ago, announced a major initiative in try to promote faculty in their engagement. A number of efforts in this are going on. I think I can only highlight two concrete steps. The first is to help faculty share the expertise at all levels, federal, state, community, and professional society.

0:07:55: In fact, my office managed the travel support. Let's say, if a faculty member got invited to testify either House or Senate committee, then we can pay for the travel for that particular effort. You know, many of our star faculty members, even though they have a large research program, but usually they don't want to use the research, and typically we're not allowed to use the research, and then to go back and then testify in Congress. We provide travel support. The second is to continue engage communications and also our academic innovation program to publicly recognize the value and our impact of our research and scholarship. Through many years of investment, I think our faculty really have taken leadership roles in the country, in our professional society, and then have done scholarship that are in leadership positions in their respective areas. Through research and engagement, we strive to bring value and impact to society through our scholarship and our service. So, that's my remarks. I promised David it should be four minutes, ten seconds...


0:09:14: So I hope you enjoy the rest of your conference and enjoy your stay in Ann Arbor. Thank you all very much.


0:09:25: Thank you, Jack. One of the traditions at Michigan is that university leadership continue to be very engaged in their units, and Jack spending the day listening to oral qualifying exams in engineering [chuckle] says so. He'll have to be going off to do that. But we'll move right into our first session on Innovative Research to Understand and Reduce Health Disparities. And I'll turn over to Keta Cowan, who's gonna be the chair of the session. I should mention, we are live-streaming, and we had quite a good crowd, in addition to the crowd here, watching on live-streaming. So welcome to anybody live-streaming, and just so everybody understands that's happening.

0:10:05: Good morning, everyone. My name is Keta Cowan and I just wanted to say... Just want to take a couple of minutes to talk about how it is I come to be here. When Dr. Delva invited me to be the moderator, I was excited because it gave me the opportunity to talk a little bit about the university's impact on inequality right here at home. Through collaborations with Dr. Delva and the School of Social Work as well as the Schools of Public Health and of Information, we've established an ID program to reach folks who are undocumented and also from marginalized communities to enable them to access fundamental goods and services including access to law enforcement. I really appreciate your work and your efforts in that regard. With no further words from me, I just wanna introduce you to Dr. Jose Bauermeister.


0:11:07: Good morning.

0:11:08: Good morning.

[background conversation]

0:11:28: Good morning. It's actually a pleasure to be here. If you know me, you know that I've lived my life in Ann Arbor, literally. A third of my life, most of my adulthood actually, has been spent here, although I'm originally from Puerto Rico. And so it really is fun to be here. I did my master's and my PhD here, thanks to the Rackham Merit Fellowship, so I'm always grateful to Rackham and to the School of Public Health and ISR for my training. We're talking about sexual inequalities.

0:12:00: I'm always conscious of sexuality, and so I decided I would use my time to talk about sexuality as one of the places where we don't do as much work as we could be doing, since a lot of the work that I do actually is inspired by many of you in the room, including Dr. Williams.

0:12:17: I got into grad school because I was interested in changing the world, as many of you in the room are. In Puerto Rico, I would look at literature on sexuality and the experience of Latino gay men and HIV, and they didn't actually speak to me. I was like, "Who are these people who are risky? Who are these people who are engaging in these type of social circumstances? And what's good about being a Latino gay man in the US? What else is there?" And of course, we've now learned the hard way that if you only recruit in STI clinics, then the only narrative that you're going to put forward is that Latino gay men are getting STIs. Well, right, so how do we reach communities? And how do we think of communities in a very different way?

0:13:06: And then I had the good fortune of getting recruited back into the faculty and Health Behavior after I did a post-doc at Columbia by Dr. Zimmerman in the Health Behavior and Health Education. And I thought about invisibility and representation again. And what started as a small initiative, actually through OVPR funds at the beginning and then through a MICHR grant, became now an official Center for Sexuality and Health Disparities, which I founded in 2009, originally called SexLab. There were no beds. I always have to say that to people. [chuckle] And has now matured to be a full center with over 10 faculty and over 40 federally-funded projects around the world. So the impact of Michigan can start with a little RMF guy who wants to make a space at Michigan to have other people do work. And even after you leave, it's still there.

0:13:58: I completely changed my talk. Sorry, Jorge. Sorry, Keta. I completely changed my talk after listening to the panel yesterday, because although we recognize inequalities in many different areas, sexuality is never there. And so, this is me and my soapbox, being like, "We need to start thinking of sexuality and removing it from race and gender, looking at it in its own way, and then bringing it back into the intersectionality space." Otherwise, we might make some assumptions that are really problematic. And I think one of the big ones is thinking about education. The most recent CDC data shows that in the past 30 days sexual minority youth are actually three times more likely to avoid going to school for fear of safety. And so, in this climate, we really have to start thinking of sexuality as a life course issue and as a thing that compounds all the other inequities we find in our society.

0:14:50: The world in changing. I don't think I would have ever seen the day, at least not as quickly as I saw it, where the NIH has actually made it into policy that sexual and gender minorities are now officially a disparity population that deserves attention in its own right. This comes out of a report that NIH had put forward through the IOM as a request to start understanding why do we see LGBTQ+, I'm gonna use alphabet soup on this one, to really understand disparities. And there are many that we don't know about. There are many that we need to know more about. But I think the one that just remains as a global challenge is HIV, and so I'll use HIV as a platform to understand sexuality inequities globally. And it goes all the way from prevention to care. I don't think of them as negatives versus positives. I think of the whole combination.

0:15:48: And so, when I started training in health behavior here, we learned intervention theories. We learned how to build programs to change vulnerability. And I ran into this report, during my first year of grad school, from the NIMH. And it had these amazing theorists, some of whom were Michigan social scientists like Marshall Becker, who had convened at a task force in 1991 to talk about why are some groups more at risk for HIV, and their conclusion was that HIV infection is, first and foremost, a consequence of behavior, that it is not who you are but what you do that determines whether or not you expose yourself to HIV, the virus that causes AIDS. And at the time, that made a lot of sense. It was a response to stigmatizing labels that perhaps were setting wrong policies and information. But as we've gotten more sophisticated to understanding inequality in this country, we started seeing some unequal patterns of distribution of disease.

0:16:46: This is a CDC dataset of people living with HIV who have been diagnosed around the country. The deeper the red, the greater the concentration of HIV cases in each county. And I've chosen to interlay that with poverty rates in the country across county. And visually, you can see the severe overlap between where HIV is concentrating. And you can change this for the Gini coefficient, you can change this for racial minority densities, and you would see similar patterns.

0:17:20: I was fortunate to train under Rafael Diaz, who actually had written this great report countering the NIMH task force, saying what the task force didn't actually get is that "who you are," quote unquote, not in terms of your individual identity but in terms of your social location, within a context of oppressive factors, determines what you can and cannot do. And so, to spend time building interventions at the individual level without thinking where people live is really a missed opportunity. And that really changed how I thought about the world. Within HIV we have this increasing racial disparity. In metanalyses, we find that black men who have sex with men actually engage in less risk behavior than their white counterparts across every socioeconomic status. And yet they are 20, 30, 40 time fold at risk of HIV acquisition in this country. One in two African-American men who have sex with men, whether they identify as gay or bisexual, will acquire HIV in the country according to the most recent CDC projections.

0:18:23: This is also true for Latinos, so men like me: One in four chance of becoming HIV-infected in our life course, with some cities, including San Juan, where I'm from, and New York, Miami, Houston, and Chicago, carrying the burden. Philly, not so far away from these trends. And so, if we really wanna do social change, we have to start thinking about behavior in the context of space, place, and location, and the way we perform sexuality within these spaces. And so, as I've struggled... I'll just say that I've struggled with thinking about, "How do I make the biggest use of my training from Michigan in the community?" I've just come to a recognition that a national blanket will not work, that we need to have local solutions for local problems. Whether it be through linkage, and prevention, and care services, whether it be retaining folks in care, and creating social services tied to HIV care services, or whether it's actually scaling of prevention through biomedical prevention, which we now call PrEP, or getting folks living with HIV to be virally suppressed, which now the CDC has come out to acknowledge that once virally suppressed, the risk of HIV transmission is zero. We call that the UequalsU: Undetectable means un-transmittable.

0:19:39: And so, to do that, I've really thought about CBPR and Community Participatory Action Research as being crucial to that endeavor in bringing inequalities solutions forward. I have been working on this little model since I was faculty here in 2009, and we continue to use this work in Metro Detroit through a community partnership that MICHR originally founded and supported, and then kind of took a life of its own, called UHIP. And now we're actually using the same model in Philadelphia. It's kind of an interesting model where we think about individual behavior as being layered by both community processes and social vulnerabilities. And so, it gives us a nice road map to think about how these different factors that have been recognised in the literature come together. I'm home, so I feel like data that I'll show you is Michigan data. This is Southeast Michigan. This is one of our initial studies where we did... Where we mapped housing vacancy and participants in a survey. We had 429 sexual minority men and transgender women participate in the study around a social vulnerability to HIV. I'm not sure if you can see it very well. Those little green dots are participants. And we've randomized them a little bit. We've scrambled them so you don't really know exactly where they live. But the deeper the purple, the more housing vacancy.

0:21:05: And so, you can see that we had a lot of participants concentrated in these areas. Since then, we've actually shown a relationship between housing vacancy, having to engage in survival sex, and HIV risk. So, using sex as a way of getting access to resources like housing, food, and shelter, among young men and trans people. And we've also tried to locate AIDS Service Organisations, or ASOs, in the context of HIV prevalence. The good news is that agencies in the Southeast Michigan area are actually located in the areas of highest density of HIV. We have modeled the distance of participants to those ASOs, and lo and behold that the nearer you are to an ASO, the more likely you've tested in your lifetime, and the less likely it is that you have a serodiscordant or a sero-unknown partner in the past 30 days.

0:21:57: That's actually really good news. It means that we can think about systems and co-locate systems where people have the greatest need. One of the plays that I've struggled with is thinking about how do we link that then to intervention? And so, one of the places we've sent millions upon millions of dollars is building these test locators. We... Choose our outcome, right? In any health condition. And if you have your ZIP code, it'll tell you the nearest clinic where you can go get a service. This is the services for HIV according to and you can see there are actually a multiple set of agencies available for both HIV testing, as well as Ryan White CARE and family planning. What we have failed to do, nationally if not globally, is to really take a step back and think, as public health practitioners, if we're actually giving the best optimal care. Can we really say without a certainty of a doubt, that no matter where we send young people, particularly young people of color who are sexual minorities, to go get tested, that we're gonna treat them fairly and equally?

0:23:00: And so, we adapted a business tool, mystery shopping, to train young people from the community to actually go and get tested. At the Michigan site, it was actually some of my MPH students and PhD students at the time. And in Philadelphia, we're now recreating this, and we're also doing it in Atlanta and Houston, and we're actually using folks from the community, not necessarily graduate students. And we wanted to understand what is the experience like when you walk into a clinic. And overall, the clinics did very well. One of the places where we found big disparities was around the quality of the test counseling experience that folks were receiving, whether they went to an HIV testing-only site, usually community-based agencies, or going to comprehensive HIV and STI testing sites where they're supposed to be receiving all sorts of other testings, for syphilis, gonorrhea, chlamydia, trichomoniasis, you name it.

0:23:55: And increasingly, one of the things that kind of struck to me was that only 43% of all sites actually offered a shopper action steps to reduce their risk. Missed opportunity one. Missed opportunity two is that they actually don't talk to them about condoms, figuring out a condom that makes sense for them, and most importantly, lubricant. If you don't talk about lubricants, you're already lost in the mechanics of man-to-man sex. I'll just give you one story 'cause I think stories are really powerful. "Initially, the nurse practitioner appeared friendly and well-meaning." This is a shopper debriefing on his experience. "She told me about the different options and testing procedures that were available to me. When I... " She then offers the penile swab as a test, which is actually not clinically indicated anymore. "When I refused to do the penile swab, she replied, quote, 'Our clients don't have the option to pick and choose which tests they can or cannot do. It's all or none here. If you don't do the penile swab, I can't offer you any other tests.'" That's actually an inaccurate statement from the nurse. "When I still refused, she replied that this was a simple, painless procedure, and that it would take a few seconds. She almost seemed disappointed that she had not intimidated me into doing the penile swab."

0:25:06: "Then she did the rapid HIV test, and did not explain anonymity or confidentiality." Problem number two. Michigan is an HIV prosecutable state, and we can talk about that in the Q&A if you want. "She asked me to wait outside, and did not do any counseling for me. She seemed really disappointed that my results were not what she expected. She said, quote, 'With your history, I was certain you'd be positive, but you seem to be negative. I'm going to give you some condoms. Use them every time for sex.' Overall, this was the worst testing site, where my autonomy was taken away and I was denied services because I refused to toe the line. This is a site that should be avoided for its unethical approaches and for being unapproachable to clients. This is definitely a site that should be avoided in its entirety." So we spend millions of dollars encouraging young people of color who are sexual minorities to get tested, but there's no real way to actually contest the fact that maybe they're not getting tested because they're being victimized in the spaces that we have assumed to be safe for them. We have built a project here in Michigan, called Get Connected, that actually tries to match mystery shopper data with participants.

0:26:06: And we actually try to say, "Can we date you? Can we help you connect your needs with the best agency in your community so that you have a really good experience?" That has now been picked up by the NIH as a full efficacy trial that is being launched in Atlanta, Houston, and Philly. We're really excited about that. And one of the things... I'm perfect on time. One of the things that we are doing is actually giving results back to the agencies. It's a technical assistance where we say, here's what participants are saying about your services. It's kind of like a Yelp but more structured and much more scientifically rigorous. And then we give them pro bono technical assistance. We don't want to punish agencies; we wanna help them do better. And so, I think we need to think more ecologically. I think we need to think about sexuality as a key in equality. But always locate it within space and place. And I look forward to your ideas and thoughts on this. And so, with that, I'll end. Thank you.


0:27:15: I'd now like to introduce Dr. Desmond Patton.


[background conversation]

0:27:39: Alright, good morning. It is wonderful to be back at the University of Michigan. I want to thank my mentor, Jorge Delva, for the invitation, and the symposia organizers. And to be a part of this esteemed panel has been amazing. I did not do my doctoral training here, but I did do the master's in Social Work program in the School of Social Work. And one of the things that has stayed with me from my time as a master's student is the importance of community and listening to the community, in research and in practice. And so, that particular perspective has been a part of my research practice and methodology since graduating. And so, for the last five years, I have been working to understand the link between social media communication and gang violence among young people who live in violent neighborhoods in Chicago. And that work was initiated when I started my first faculty position at the School of Social Work here at Michigan. And as a qualitative social work researcher, I'm interested in the narratives that are told digitally, and specifically, how young people use language and social media as a tool to express daily life. And for youth in Chicago, those lived experiences that they express may include trauma, threats, and exposure to violence.

0:29:10: And so, to fully capture those narratives, and to move from individual-level perspectives to a population-level understanding, I've been working with computer scientists at Columbia University to create a prototype computational system that detects aggression and loss in a small Twitter dataset. And I'm gonna talk more about why we're using a small dataset, or a thick data approach, as opposed to a big data approach, initially. Today, I'm gonna talk about our multidisciplinary process, report some of those findings, and discuss some of our ongoing work moving forward. And so, when I left the University of Michigan in 2006, I went on to the University of Chicago to do my doctoral training. And there, I carried with me this focus and interest in community.

0:30:00: And so, I did a dissertation that looked at how young African-American men navigate violence and stay connected to school. But as I was chatting with them about the communities in which they have to navigate, one community that I did not consider was social media. And so, I learned a tremendous amount from these young African-American men about the things they have to do and say, and the ways in which they talk to each other on social media that might lead to offline violence, as well. And so, in my work, I focus on the city of Chicago. And many of you are probably quite familiar with what's happening in Chicago, because it's constantly in the news. Every day, every week, you may hear a new story about the uptick in violence. In the past two years, Chicago has seen a dramatic uptick in violence of around 58%. And this year, they're on par with similar numbers from 2016. And at the same time, we also are seeing a tremendous, a ubiquitous, use of social media among young people across the country. And young people are using social media to do a number of things: To find community, to learn about the news, and to share news, to find relationships, to talk about love, and happiness, and joy, and to be connected to one another. But when young people live in a local ecology of violence, that deeply personal connection, those networks, those comments, may also be embedded in the violence that they are exposed to and have to go through on a day-to-day basis.

0:31:32: And so, this work is really a continuation of my interests in the ways in which young people navigate violence. And what we're learning is that everyone is beginning to realize that there's something to what's happening on social media. And so, we have been framing the ways in which we talk about social media in very negative ways. And the Chicago Police Department has had a lotta input on this. Here is the former superintendent, who says that social media is the new way of taunting and challenging other gangs. And so, these potential threats are things that people are now beginning to zoom in on and to focus on very concretely. However, and so, our science has advanced in how we think about violence. And we now treat it as a communicable disease and situated in the most vulnerable communities of color. However, the proliferation of social media platforms has also shifted on what we can now call community. However, our approaches to gang violence prevention have not shifted online. And so, there's a critical gap in how we're thinking about violence prevention. And so, to fill this gap, I have led a set of qualitative and observational studies to more fully unpack the role of social media in gang violence.

0:32:51: And so, first, I introduced the term "internet banging" to define it as "a cultural phenomenon that has evolved from increased participation with social media that represents an adaptive structuration or new and unintended use of existing online social media platforms." And so, we examined the role of hip-hop in the development of internet banging, and highlight the changing roles of hip-hop and computer-mediated communication as a social representation of life in violent communities. Building on this internet banging thesis, we then identified Twitter communication that mimics gang behavior. Specifically, in a group conflict, reciprocity, and status-seeking. And we explore how cultures translated through online Twitter and find salient scripts of reciprocal violence within local social media networks as it is written online in real time. Youth writing those scripts anticipate, direct, historicize, and mourn neighborhood violence in their communication with one another.

0:33:57: And then in qualitative interviews with gang-involved youth in Chicago, we found that content perceived as threatening is based on the context of the post and the relationship that the reader has with the author of that post. There are several theories that have helped to frame how we are approaching this work. We know about the socially-embedded nature of identity that has been recognized by social scientists as a product of our social interactions. In a network era, however, those social interactions are rooted in offline context, and requires knowledge of that context in order to be interpreted. Contagion of Violence Theory tells us that violence spreads through networks and face-to-face interactions. Social media content and expression is rooted in offline behavior. To what extent is social media a vector for community violence? And so, that means, if you follow what we may call the "code of the street," the ways in which young people command respect and posture in the neighborhood, that may also end up online. And because youth spend so much time online, it may be hard to distinguish offline behavior from online communication.

0:35:11: And so, we've learned that there's a critical need to unpack culture and context when reviewing social media posts, that those offline events and experiences actually shape those posts, and that emojis carry a lot of weight and actually should be treated as another form of language, and that it's important to move from examining just individual users to think about what a community is saying about their context and their experiences with violence. And so, in order to look at large amounts of social media data, I realized that I needed an approach that allowed me to look at social media data quickly and automatically. But once I got involved in the social media content of young folks from Chicago, I quickly realized that I didn't know what the hell young people were saying online.


0:36:00: And so, if I really wanted to do this well, I needed to take a minute, and pause, and think very complexly about the context of the language, and to spend time reviewing a very small amount of posts first before trying it out on a big dataset. And so, from this, what we have decided to do is to take what we've learned about offline features and characteristics from the qualitative work, and then to partner with computer scientists at Columbia to then create algorithms that are trained to think like the young people from the context. I'll talk more about that process. And so, I'm gonna talk about a case study that we've done with one Twitter user and individuals in her Twitter network. The aim for this case study, computational study, was to identify offline characteristics in Twitter communication to develop an automatic coding process and to evaluate the use of natural language processing as a tool for analyzing social media data.

0:37:00: I wanna introduce you to Gakirah Barnes, and this is a young woman that I came in contact with while I was on faculty here at the University of Michigan. And what was interesting is that I was sitting at my desk and this headline came crashing across my screen that said, "The gun-toting gang girl of Chicago." And of course, I was rather dismayed and confused by the sensational title, but once I started to dig deeper, I realized that there was something more to the story. And so, the story talked about a young woman, 17 years old, that was shot and killed by a rival gang. And they talked about her life as being the shooter, or hitter, in the gang, which is a unique position for a young woman. It's not abnormal for a young woman to be a part of gangs, but to be the shooter, or hitter, in the gang is a unique position. In addition to having this prominence in a South Side gang in Chicago, she also had quite the Twitter following. By the time she was 17 years old, she had about 27,000 tweets and around 5,000 followers which placed her in the 90th percentile of Twitter users.

0:38:00: And so, in this case study, we decided to focus on events that might be related to potential retaliation due to social media communication. So we focused on two events. One: March 27, 2014, Gakirah's best friend, Raason "Little B" Patterson was shot and killed by the Chicago police. And on that day, this gun-toting gang girl was actually quite vulnerable, and grieving, and mourning the loss of her great friend, which was quite the contradiction to what the narratives we were told about her in the news. And then, two weeks later, she was shot and killed by a rival gang. So we wanted to understand to what extent would individuals in her Twitter network respond to that violence. And so, we looked at around a 34-day period and initially pulled 2,000 tweets during that time, but reduced that dataset to 800 tweets once we cleared out the noise of tweets that were not associated with those events.

0:39:02: We also realized that in order to do this well we needed a interdisciplinary process that includes multiple voices. Again, something that I learned at the University of Michigan School of Social Work. And so we decided to center the voices of young people in our research process. And so, we hired young African-American men from Chicago that live in the same neighborhood as Gakirah to be our experts on translating and interpreting data. And so, we reached out to two organizations in the city to be able to hire these young men, and they are instrumental in being able to unpack culture context in the Twitter communication, and they are considered our domain experts. And then our social work students at Columbia University are then looking for themes and patterns in the Twitter data. And so, essentially the two groups, the youth and the social work students, are providing the training data that is then handed off to our computer science colleagues.

0:40:06: Here is an example of the process that we engaged in. We start with the social media data set. We have two groups of annotations that are happening, with our community members and our social work students. We are provided a labeled dataset that is then the key for unlocking the NOP computational system. We realized that we were trying to find different methodologies and strategies for analyzing the data, and we realized it was really hard to find a qualitative system or any kind of system that would allow us to look at Twitter data and to contextualize it in very meaningful ways. And so, we just created one. With some engineering students at Columbia, we developed an annotation system for looking at tweets that provides the actual tweet and gives our annotators opportunities to first provide a description of the tweet without doing any deeper dives, and then to also assign a threat assessment to that tweet without looking more deeply into it. And then they're asked to go back with a set of cues to look at the data more complexly.

0:41:12: And so, I'll start off with the domain experts, and then I'll walk you through what our social work students are also doing, as well. And so, we asked our community experts to look at tweets and to give us input on what's happening in the tweet in their own words. Let's look at a tweet. This one says, "We lost them savages. Real in word. Only real in word left. Fuck all those others and your block." And so, on the face, this tweet might appear to be aggressive or threatening, but when we ask our young people to explain to us what's happening in the tweet, they have a different opinion. One of coders says, "Well, my first impression is that the person has a gun and a blunt of weed, and I can't take this person serious. My thoughts is that he is just trying to be cool but he is betraying himself by doing so."

0:42:00: And another coder says, "He lost members of his gang that he clearly respected and is now mad, so he expresses disrespect towards rival gangs, and members, and their turf." This is a very different way of approaching the tweet that could initially be categorized or deemed threatening by law enforcement, per se. Our social work annotators are looking at 50 tweets at one time, and they are overlapping with 10 tweets at any one time. Any time there is a disagreement in their coding process, that tweet is kicked back in the system, in our annotation system, to our community experts. And so, any time we are disagreeing on how we're labeling data, the final say is given to the young people from Chicago on how we should be making meaning of that particular post. Here's an example of what we do once we have set aside a set of tweets to analyze more deeply. And so, in this particular example, we have Kida smoking, thinking about Little B. And so, on the face, this tweet might look like it's talking about someone using a substance, but when we consult with our community members and do a deeper dive, we realize that Kida is actually a name of an individual that was shot and killed from an opposing gang.

0:43:20: And so, smoking Kida is actually in reference to smoking an individual from a rival gang, and they're using this to also reference a friend that was also shot and killed. And so, this could be easily misunderstood as a post that's just about substance use, when it could be a potential post for retaliation, as well. Another thing that we also do when we're applying annotation is, when we're looking at our annotation process, is to pull out themes and patterns that we think are really important. So here, Gakirah Barnes says, "Don't get caught on the 800 block, lame ass. You better take that shit on the stormy spot." And so here, we pull out things like "800 block" and "stormy spot" and ask, "Well, what is the significance of these words?" In this particular post, what we've learned is that Gakirah's actually telling us that there is a gang boundary that should not be crossed, and that individuals who are reading this post should be aware of this gang boundary. And so, what we've learned thus far from observational work is that the things, the experiences, the exposure to the violence that happens in the community, become the very things that are talked about and communicated online. And we found two important themes in our observational work: Themes of aggression, these are mentions of gang affiliation that happen on Twitter, gang activity, direct and indirect threats, and mentions of birthdays, and mentions of jail and prison time. And also expressions of loss.

0:44:46: We also have a category of other which includes all the other amazing things that happens in people's life that they also communicate on social media. I'm gonna go through this pretty quickly 'cause I'm running out of time. And so, we take those offline features and characteristics; they become the very thing that our data science team uses to predict and detect content on social media. And so, they take that aggression theme and loss theme as their binary classifications. And so, when working with the computer science folks, they have these amazing tools at their disposal. One of the things that they use is a part of speech tagger. And so, they had all these gold standard part of speech taggers that normally work on the English language. They quickly realized that their part of speech taggers do not work on this particular context because it's just not prepared for this particular context. And so, for example, the gold standard part of speech taggers did not understand what "lame ass" meant within this particular context, and so it did not work. And so, we had to create a new part of speech tagger that was trained to think like young people from Chicago. And when we did this, we realized that the part of speech tagger was then able to identify aggression and loss in the data with 81.5% accuracy.

0:46:05: We then used the part of speech tagger and the other binary classifications to then classify the data as either aggression, loss, or other. And when we use an f-measure, which is a measure of our ability to be able to classify that data, we only get around 62%. What we're realizing is that, again, our ability to annotate this data and to classify this data, to get the true meaning of a data, still remains important. And being able to talk with community-based organizations about how we should interpret this data remains an important piece. And so, one of the things that we're doing is trying to replicate this study with additional offline features, and to be able to partner with community-based organizations that can actually rule out these classifications in real time. And so, now, we're replicating the case study at this time with more users, more tweets, and also adding in images as a part of our analysis, as well. This is an example of the image analysis that we do. We draw bounding boxes around pictures, to really try to understand what is happening in the picture, to use it as a part of the NOP process, as well.

0:47:11: Again, this is all challenging in that right now our database is just of black and brown youth. And so, we need to be able to have a system that incorporates other types of folks in the image, as well, so that we don't create an NOP system that's just looking for black and brown youth as threatening or violent in the system, as well. And so, to end, we currently have a study under review in which we are hoping to be able to partner with Cure Violence to be able to provide them with realtime NOP classifications. The goal is to be able to create an alert user-face that then spits out these classifications on aggression or loss, sends a text message to a violence interrupter in Chicago that would then have information not about the individual but about a block, or a community, or a street-level interaction that is unfolding on Twitter. And then we will hope that Cure Violence can take that information and be able to develop targeted interruption strategies and digital education that they can then use to hopefully reduce offline firearm violence. Thank you.


0:48:20: I'd like to introduce Dr. David Williams.


[background conversation]

0:48:45: Okay. It's truly a pleasure to be here at the University of Michigan. I have spent 22 years of my life in the state of Michigan, 18 of those at the University of Michigan. I've lived in Michigan longer than I've lived anywhere else, and it's always an honor and pleasure to come back home. So much so, I was here at this podium last week [laughter] speaking for James Jackson's symposium.

0:49:08: What I wanna do in the time that I have today is to give you a high-level overview of the challenges of inequalities in health in the United States and some areas of opportunity in order to effectively address them. I was trained as a sociologist here, and I learned that socioeconomic status, was measured by income, education, occupational status, is one of the most profound determinants of variations in any desirable resource in a society. If we take, for example, the SAT test, the Scholastic Aptitude Test, that some are calling the Student Affluence Test because of the powerful relationship between SAT scores and family income. This is national data for the US in 2014, and you can see that every higher level of household income is associated with a higher SAT score. This raises profound questions of what these tests capture, how we should use them. But maybe we can come back to that in the discussion. What is true of SAT scores is also true of health outcomes in the US.

0:50:22: Here is data from the Panel Study of Income Dynamics looking at all-cause mortality risk by household income. And you can see, every higher level of household income is associated with a lower risk of overall mortality, and with low-income Americans having rates... Relative risk of overall mortality that is three times higher than that of high-income Americans. Well, those data, the power of SES, would prompt us to think there would be large racial/ethnic differences in health because of the association between race, ethnicity, and socioeconomic status. If you look at infant mortality, we see even with the earliest health data, the beginning of life, you can see the risk of dying before one's birthday is patterned by race with African-Americans having especially elevated levels of infant mortality compared to whites, and Native Americans also have an elevated risk. The data on Hispanics and Asians are more complicated than these patterns show. Immigrants of all racial/ethnic groups have better health than their native-born counterparts, but with increase in length of stay and generational status, the health status of immigrants worsened.

0:51:37: These racial ethnic inequalities not only exist today, but they are quite persistent over time. If we look at 1950 to the present, we have data only on blacks and whites, and you could see the... There's good news and bad news in these data. On the one hand, and you can see the increasing levels of life expectancy for both African-Americans and whites. That's good news. You can see a narrowing of the overall gap: An eight-year gap in 1950 compared to a four-year gap in 2010. At the same time, a four-year gap in life expectancy is quite large. If we froze the life expectancy of whites, and had a life expectancy of African-Americans increased at the average rate at which life expectancy has increased in the last decade, it will take about 28 years to close the four-year black/white gap. You can see these numbers in the table. Look at the life expectancy of whites in 1950: 69.1 years. How long did it take for African-Americans to catch up to the health that whites had in 1950? You'd see it wasn't until 1990, 40 years later, that blacks equaled the health that whites had in 1950. When I started my career, most researchers thought that racial/ethnic differences in health was simply a function of socioeconomic status. We now know that life is much more complicated than that.

0:52:55: Here is national data for the United States: Life expectancy at age 25. And at the age 25, the average white person lives five years longer than the average African-American. However, race matters at every level of socioeconomic status. And in fact, there are large socioeconomic status within each race. The gap in health between whites with a college degree and whites who have not finished high school is 6.4 years; bigger than the black/white gap. And within African-Americans, there's a 5.3-year gap. So socioeconomic inequalities even larger than racial ethnic ones. At the same time, race matters for health at every level of socioeconomic status. White high school dropouts live 3.1 years longer than black high school dropouts. And the difference is even larger among the college-educated. And one of the most stunning statistics I will show you today, national data for the United States, that the best of African-Americans, those with a college degree or more education, have lower life expectancy than their similarly educated white counterparts, than whites with some college education, and even lower life expectancy than whites with a high school completion.

0:54:03: Why does race still matter so much for health, even after we've taken social-economic status into account? I and others have been involved in answering the question: "Could racism be a critical missing piece of the puzzle to understand the patterning of racial disparities in health?" I think it's important to understand the distinction between individual and institutional racism, so let me give you an example. Some researchers at Portland State University asked a simple question: "When a black person and a white person stands at a crosswalk intending to cross the street, does your race determine how long it takes to cross the street?" And they found that multiple cars were twice as likely to pass a black pedestrian waiting to cross the street, and that blacks had to wait 32% longer than whites to cross the street.

0:54:48: That's individual discrimination. We're looking at the individual behavior of drivers as they see a pedestrian waiting to cross the street. But let's think of a mechanism of institutional discrimination. I'm gonna pick one linked to waiting. In the 2012 presidential election, African-Americans waited, on average, almost twice as long to vote as whites did. And you can see Latinos waited 19 minutes compared to the 12 minutes whites waited. Now, none of this reflected individual behavior of precinct workers in terms of how they treated persons based on different races. Instead, it reflected a number of institutional mechanisms, linked to budgeting and space constraints, and local administrative procedures that determine the number of persons served at a particular polling site, and how many precincts were in a given area, and what's the level of staffing. Although there were no individual discrimination, there was still systematic differences in how long it took individuals to vote.

0:55:49: One of the most powerful mechanisms of institutional racism in the United States is residential segregation. It's a striking legacy of racism, and it has pervasive effects on health. Today in public health, researchers say that your zip code is a stronger predictor of how long and how well you will live than your genetic code, because where you live determines your access to educational opportunity, to employment opportunities, to quality of neighborhood and housing environments, to whether it's easy or difficult to live a healthy lifestyle, to exposure to toxins, even access to quality medical care, and quality of city services. A broad range of factors are closely tied to place.

0:56:33: Two of America's most eminent sociologists, William Julius Wilson and Robert Sampson, looking at 171 largest cities in the United States, says there's not even one city where whites live under equal conditions to blacks. And the worst urban context in which Whites reside is considerably better than the average context of black communities. And in fact, even the racial differences in socioeconomic status, we so are familiar with, didn't just come out of nowhere; they reflect the successful implementation of institutional policies. David Cutler, until recently the Dean of the Social Sciences at Harvard, a leading economist in this country, using fancy econometric models I cannot even fully describe, but isolates the impact of residential segregation in a national study of blacks and whites, and shows that if you could statistically eliminate segregation, you would completely erase black/white differences in income, education, and unemployment, and reduce black/white differences in single motherhood by two-third. And so, these large racial differences in socioeconomic status that we see... And here, for example, is data on 2015 income in the United States, and I'm just translating it in a way you can't possibly miss the point. For every dollar of household income that whites receive, Asian households receive $1.23. Asian households have more persons contributing to household income than any other group.

0:57:56: A per capita measure of income would lead whites to have the highest level. But for every dollar of income that white households receive, Latino households have $0.72; American-Indian households, $0.62; and African-American households, $0.59. What's striking about the $0.59 figure for African-Americans? That's identical to the racial gap in income in 1978. I did not misspeak. In 1978, blacks earned $0.59 for every dollar whites earned, and in 2015, Blacks still earned $0.59 for every dollar whites earned. And as bad as the data on income looks, it dramatically understates the racial differences in economic circumstances, because income captures only the flow of resources into the household, it tells us nothing about the economic reserves that household has to cushion shortfalls of income. We get that by looking at wealth. And in 2011, for every dollar of wealth white households have, black households have six pennies and Latino households have seven pennies.

0:58:58: Racial inequities in socioeconomic status are not random events, are not acts of God; they reflect the successful implementation of social policy, and they powerfully illustrate the ways in which institutional mechanisms of racism is produced in a truly rigged system in the United States. And then, there's also individual discrimination at the interpersonal level that has been documented in so many domains of life. I and others have been interested in what impact does the subjective experience of discrimination have on health, and could the subjective experience of discrimination be one type of stressful life experience that has pervasive negative effects on health? To illustrate the work in this area, I'll show you one scale that I was instrumental in developing, called the Everyday Discrimination Scale, that captures the day-to-day indignities linked to discrimination, treated with less courtesy, less respect, receiving poorer service than others. It was fielded for the first time when James Jackson and I led the Detroit Area Study back in 1995, another Michigan Resource that facilitated this research. But to illustrate this, I'm showing the work of another Michigan alum, Tene Lewis.

1:00:12: Each line on this slide represents a different empirical study in all of the studies looking at the impact of everyday discrimination. Higher levels of everyday discrimination predicts higher levels of coronary artery calcification for individuals followed over five years, predicts higher levels of inflammation, higher levels of blood pressure. Pregnant women who report discrimination give birth to lower birth weight infants. The study of the elderly followed over time: High everyday discrimination predicts more rapid declines in cognitive function, high everyday discrimination predicts poor sleep, is an independent predictor of premature mortality, and is a predictor of visceral fat. So, across a broad range of outcomes, and this is just one example to illustrate the power of discrimination. There is also an impressive body of work documenting that discrimination also matters for minorities in the areas of medical care. I served on the IOM Unequal Treatment report committee that released the report back in 2002.

1:01:09: And so, I wanna shift in the few minutes I have left to talking about what can we do about these problems and what are the areas of intervention. First, we need medical care that addresses the social context. The World Health Organization asked, "What do we accomplish if all we do is to treat illness and send people back to live in the same conditions that made them sick in the first place?" How can we take the social challenges that individuals have and use the healthcare context as a place to link individuals to other resources? There're many examples of programs that do this. I'll show you one of them. The Medical-Legal Partnership born at the Boston Medical Center. In the Pediatrics Department of the Boston Medical Center, primary care provider can refer to a number of specialists. One of the specialists they can refer to are lawyers. Yes, the hospital has on-site lawyers to solve the problems in the lives of their patients.

1:02:01: Because if a mother comes to the Pediatrics Unit with a child who has asthma, and the asthma is secondary to the mold that is faced in the apartment, all the asthma medication in the world will not help that child to breathe symptom-free if the child goes back to live in the same conditions. And the mother who has had no success in getting the landlord to address the problem, it does make a difference when a lawyer calls and says, "You're in violation of the housing code of Massachusetts and we will sue you if you don't fix the problem."

1:02:35: What else do we need to do? We need to start early. I'll give you an example of the impact of starting early. The Abecedarian project in Chapel Hill, North Carolina, that randomized poor children, 80% of them African-American, at birth, to an early childhood enhancement program. And for the first five years of their life they were in that program. By their mid 30s, those who got the program have lower levels of multiple risk factors of cardiovascular disease and metabolic disease. Here, for example, are the differences in systolic blood pressure between the treatment group and the control group in their mid 30s, linked to what they were exposed to, birth through five. What else do we need to do? We need to address economic wellbeing. Yes, economic policy is health policy. Improving the economic circumstances of individuals has a dramatic impact on health. Nationally, if we look at data, when the black/white gap in health narrowed dramatically between 1968 and 1978, as a result of the gains of the Civil Rights Movement and the Anti-Poverty Poverty, is one window of time where we had the health of African-Americans improving more rapidly than the health of whites, and a narrowing of the black/white gap.

1:03:50: There are a number of resources that provide a road map of how we can cut poverty in half in 10 years, and we need to think of these not only as economic strategies but as health strategies. We also need to improve neighborhood and housing environments, because all of these powerfully drive health, and we have high-quality evidence from the Moving to Opportunity programs and a range of other outcomes and other studies that show simply changing neighborhood environments, doing nothing else, changing the neighborhood environment can lead to improved health 10-15 years later.

1:04:27: We also need to undo racism in the United States. And when we think of undoing racism, we need to think of it very comprehensively. And importantly, I'm emphasizing that it's not enough to open the doors of opportunity; we need to ensure that individuals are able to go through those doors. I wanna illustrate this point by looking at the success of affirmative action for women in getting into medical school in the United States. In 1965, 6.9% of graduates of medical school were female; in 2010, 48%; today, it's 50%: Dramatic improvement in the number of women. Affirmative action also was designed to help minorities. You could see, in 1960, about 3% of medical school graduates were African-American, and in 2010, only about 6.5% are African-American and 6.7% are Latino. Why have women been so much more successful in taking advantage of the opportunity of affirmative action than minorities? Well, because women were more prepared to walk through those doors when the doors were opened. It's not enough to put groups at the starting line. People need to be prepared to be able to effectively run the race, and we need to think of what other shackles that we have to move from their legs so that they can be successful.

1:05:51: What is the challenge we face in the United States? A report from AAMC in 2015 reported that in 2014, there were 27 fewer black males in the first year of medical school than there had been in 1978. In the mid 1960s, 2.9% of all practicing physicians in the US were black, and in 2012, only 3.8% of all practicing physicians in the US were black. 5.2% were Hispanic. Why is this important? Because we know, research shows, that persons from underrepresented minority backgrounds are much more likely to practice in primary care specialties, much more likely to work in underserved communities. And as a nation, the University of Michigan... And I'm thankful to the University of Michigan for its affirmative action program, the minority fellowship that enabled me to matriculate here. Without that, I wouldn't be standing here today. But we need to still remind ourselves, in the words of Plato, that's there's nothing so unfair as the equal treatment of unequal people. And I've showed you, race is a powerful predictor of inequality, and there's nothing so unfair that adherence to a notion of equal treatment of people who are fundamentally unequal.

1:07:04: Before I close, I wanna mention that one of the things we need to do to improve health is to keep the safety net in place. Weakening the social safety net has a disproportionate negative effect on marginalized groups. Let's take a quick walk down memory lane. The 1981 Omnibus Reconciliation Act, early in the Reagan administration, led 500,000 people to lose welfare, a million people were dropped from food stumps, 600,000 people lost Medicaid, the funding cuts led to 250 community health centers closed around United States, a million children lost reduced-price school meals, and the WIC program could only serve a third of those eligible. Research then documented pervasive negative effects on health in pregnant women, in children, increases in infant mortality in poor areas of 20 states, preventable childhood diseases, rise in poor populations; there were a range of negative effects. We are in an era where there are proposals being made to similarly weaken the safety net, and we have to work together to address it.

1:08:10: If we have all of this evidence, what is holding us back? What's the elephant in the room? Why we aren't making more progress in addressing these inequities in health? I go back to WEB Du Bois in 1899, published a book on "The Philadelphia Negro," has a chapter on Negro health in that book. And in the chapter on Negro health, he says, "The most difficult social problem in the matter of Negro health is the peculiar attitude of the nation toward the wellbeing of the race." And Du Bois continued: "There have been few other cases in the history of civilized peoples where human suffering has been viewed with such peculiar indifference." That peculiar indifference, that lack of empathy, that unwillingness to take the necessary political steps that need to be take to address the problem of inequities, many scholars today are studying and calling it the empathy gap. We need to build the political will, is our biggest challenge to address the problems that we face.

1:09:11: Let me conclude. Racism in its multiple forms is alive and well. It has powerful effects through policies and procedures that are deeply embedded in social institutions. We need to acknowledge and understand the current manifestations of racism. We need redoubled efforts to mitigate its pathogenic effects. And we need to create the political will and support to dismantle the social structures that support racism, ethnocentrism, anti-immigrant sentiments, and all forms of injustice and inequality. And I leave you with the words of Martin Luther King: "True compassion is more than flinging a coin to a beggar. It understands that an edifice which produces beggars needs restructuring." We live in a society that has many edifices that is producing inequality, and we need to work together to restructure them. Thank you.


1:10:12: If we can have our presenters come sit upfront, we'll begin a discussion and take questions.

[background conversation]

1:10:51: Okay. Do we have any questions for our presenters? Yes, please.

[background conversation]

1:11:04: Question would be for Professor Patton. And to all of you, I really appreciated your presentations. But I'm curious about the technology or methodology that you've put together to identify these gang-related texting and Twitter posts, if it's valuable at all to law enforcement. Have they been contacting you and asking for your help in using that type of methodology in their own services?

1:11:38: Great question. Law enforcement has been using social media surveillance for quite some time. And the ways in which they use social media surveillance is oftentimes for evidence and used in a very punitive nature. And so, we're hoping to craft a methodology that produces a more holistic way of thinking about how young people are expressing themselves, in a way that is more connected and true to the actual meaning. Now, being able to predict what someone is actually communicating online is a rather impossible thing to do, 'cause you would really need to be in someone's head to actually know what they're trying to say. But through our practices, we are hoping to be able to unpack closer meaning. We have struggled with how to work with law enforcement in this space. We have different approaches, and we're interested in very different things. I'm interested in keeping young people out of the criminal justice system. And so, we are hoping that through partnerships with community-based organizations that they can then identify content in social media that they should actually act on, and they can make decisions on whether or not to include law enforcement in those decisions.


1:13:00: Alright, good morning. Whoa. [chuckle] Enrique Neblett, PhD, 2006 in Clinical Psychology. Thank you very much for your wonderful presentations. My question is for Dr. Williams. I am writing a commentary right now on challenges in the field, in the area of racism and health, for psychologists. And one of the things that I've really been struck by is that a lot of the things that we were saying are problems, in terms of... That need to be addressed, are not new things. They're things that leading scholars in the field have been saying for a very long time. And so, I really appreciated your point about what [chuckle] the issue is and what needs to change. Yesterday, another presenter, Odis, made a point about how sometimes we may need to change the ways in which we get the message across. Chief Justice Roberts made a comment about research being, I think, being gobbledygook. And so I wonder... My question is, what recommendations do you have, given that there is this sorta attitude: We know what the problem is, we know [chuckle] some of the things that need to change, but the changes aren't happening. As researchers, and scholars, practitioners, all the folks who are at the table, what can we do to address that problem?

[background conversation]

1:16:47: In a way that their minds are open to listen to what we have to say. So I think it's a priority for research.

1:17:02: Again, thank you all for your presentations. They were very, very powerful. And with that, I'd like to follow up on the previous speaker's comment, and as something very practical, and that is... Starting with you, David. You know I typically agree with everything you say [laughter] but the idea of empathy and political will, to me that was really a very powerful thought, particularly in this climate that we're in now. And I just wondered, from each of you, have you had an opportunity to present some of the work that you're doing before politicians or other decision-makers? And do you have any examples, even at a small scale, where things have actually worked, that they heard what you said? Because the presentations are all very powerful in terms of what needs to be done, but getting that across, in terms of the previous question, the communication, that all makes perfect sense, but I just wondered if you had any specific examples that you could draw on for us?

[background conversation]

1:18:10: Alright. I can talk about two sets of experiences why I think it's... How we talk about things is important. I served as a staff director for the Robert Wood Johnson Foundation's Commission to Build a Healthier America. That was a national, bipartisan commission, with prominent Democrats and prominent Republicans, to look at the challenges of America's health. Why is America doing poorly in terms of health in general, and why are there large gaps in health by race, by socioeconomic status, and by place? And, I mean, most of the commissioners, were not health experts, but we had prominent people like former Senate Majority Leader Bill Frist, for example, had a former governor of Indiana as a commissioner, so, I think, fairly prominent Republicans as well as prominent Democrats. And I think one of the things we did... This is an example of putting that to work. What we were doing was raising awareness levels of the social determinants of health. We did, before the commission started, research with the leading policy-makers, on both the House and Senate side, the key committees that deal with policies that affect health, on both the House and Senate side and within the Executive branch, about the area in which this commission was going to do work. And we found that not one person we spoke to knew what the social determinants of health was.

1:19:47: And that's language that we talk all about: Importance of addressing the social determinants of health. Nobody knew what that was. And we did work with communications experts, and instead we came up with language. Instead of use same social determinants of health, it said where you live, learn, work, and play, and the opportunities to be healthy in those places have more to do with health than going to the doctor. People get that. So it's an example of... But I'll also give another example, from the communications work we did then, was we did communications work looking at both Democrats and Republicans, how they respond to certain kinds of messages. And we found, for example, just to give you one example, that many progressive people will talk about one of the challenges we need to address is to level the playing field. And we found that language about leveling the playing field turns Republicans off. But you can get both Republicans and Democrats on the same page if you talk about we need to create opportunity for all, and finding strategies to increase opportunities for all. It's the same thing, but it's how you say it can... And it's just one example of how you approach topics become important.

1:21:03: One other quick example I would give is all of the policy experts we spoke to, every Democrat, every Republican, said that any recommendations we made in this space about improving America's health should take personal responsibility into account. Every Democrat we spoke to said you have to take personal responsibility into account. Many people who do work in the social determinants of health don't ever talk about social personal responsibility. And in fact, we argued it's two sides of the coin. Yes, the individuals need to make good choices, but there's a social responsibility to remove the barriers, and to create the opportunities, and that many Americans live in places where a good choice is not even available, and we need to create that. One practical example I did years ago was commissioned by the Kansas Department of Health: To produce a report on disparities in the state of Kansas. It was just at a time when the state of Kansas was voting to create an Office of Minority Health. And I had a chance to brief the House and Senate committees of the state of Kansas on the report. And I identified all the domains of the social determinants and where action was needed, and I specified actions that were needed, what the government needed to do, what the business sector needed to do, what community organizations needed to do, what individuals needed to do.

1:22:33: But in my presentation to the legislators in the state of Kansas, in every domain, the first thing I mentioned was what individuals needed to do, because I knew who my audience was. If I started with what government needed to do, I would lose them. But now that I have come on the same page with them to talk that there are things that individuals can do to improve their health, and then go on to talk about all the other levels at which action was necessary, they were nodding their heads as I spoke, because I first built a bridge of communication with them. And I think those are some of the strategies that we need to implement: Understand where people are coming from, and try to find common ground, and once you've established that common ground, try to push them and their minds in a new direction to show them examples of other opportunities that can have even a greater impact.

1:23:29: I've had several conversations with various federal and state-level law enforcement agencies, and one of the things that I do when I'm chatting with them is to have them unpack social media posts. And they all have an opinion on what that post means, based on their experience and connections in various communities. And then I will layer in additional offline context. And so, I'll add, "Well, did you know that this post actually meant grief?" Or, "Did you know that this emoji actually was an indicator of grief?" And once I added those additional contextual features, it kinda blows up their initial assessment of what that post meant. And as a result, some of those practices have made their way to the Sheriff's Office in Chicago, and they have created a special task force that is dedicated to looking at social media communication and gang violence prevention. However, they are now including and centering the voices of community members and community-based organizations when developing that task force.

1:24:37: Oh boy. We have tried. I feel like sexuality disparities tend to be a little bit further behind than other conversations. In the state of Michigan, we provided data for the civil rights office of the state to think about the challenges of not including LGBT, by sexual orientation and gender identity and expression protections in the workforce as being one of the big problems, in our own data we've actually seen that not only is it tied to self-rated poor health if you have experienced sexuality-related discrimination in the workplace when you're a young person, they're also less likely to get a job to begin with. And so, they engage in informal systems, whether it be drug trade, survival sex, sex exchange, etcetera. And so, we have to kind of put these worlds together. All that being said, I've actually been quite mesmerized by some of the social psychology literature that's been coming out of Greg Herek's group at UC Davis. He's been tracking nationally representative samples around the country. This is... Our society's understanding of sexual prejudice, how we think about sexual minorities. And in his forecast, up until 1995, the idea of same-sex marriage was almost unattainable.

1:25:58: That... A projection that it was gonna take another 40, 50 years for that to have enough movement in the needle. And he recently wrote this wonderful report where he said something changed. We stopped thinking about the measurement of the contact hypothesis as a predictor of prejudice: Do you know somebody who's gay, bi-sexual, queer, or same-sex minority? And we started adding questions about values, what we value in a society. And that actually was the turning point on the civil side to actually get marriage equality movement to gather a lot of traction. It was no longer about knowing someone who was gay. Hi, now you do. But also, that we all equated... We had something... Equal. This idea of, like, wanting to be equal, wanting to be able to have love, being protected in your family. That... And income tax, of course, which was the Obergefell decision. We should all pay taxes and inheritance. Those things actually catalyzed, and it's a good example of how we made really quick change by integrating different theories of social science together. Yeah. There's still a lot of more work to be done, for sure. Housing, we have no protection for housing. We have no protections in the state, still, for sexual minorities. And we have criminalized HIV laws.

1:27:20: I'll just say one more thing. Sorry. I'll make a note of caution, which is the "it gets better" language that's been used to understand sexual disparities with young people. I think it is problematic, and several of us have been doing literature reviews and studies about what it means that it gets better. It's a band-aid. It does not actually give voice to the pain that a lot of young people experience. And more than that, I think, it misses on interventions at the population level. We talk... When we think about young people, we think about future expectations. You stay a lot, right? But which is what the "it gets better" is about, like, "Stay alive so that you can then have a wonderful life once you're out of this mess in your society." But there's a little crinkle in that theory on hope [chuckle] hope and health, which is the will and the way. You can have the will, the desire to want to have a future, but if you don't have a way of the... Structurally, there is actually no way of making that happen.

1:28:22: Future expectation [1:28:23] ____ to be damaging. And we've done this in a national study of young men who have same-sex attracted. The idea of parenthood: I wanna be a dad in the future. Well, if you just look at the data, yeah, it's kinda looks like it's protective. It's better self-esteem, less depressive symptoms. When you geo-locate, then, by state policies where those kids live, if you live in a state where same-sex adoption was forbidden or where the marriage ban was in place, they actually are worst off in their mental health. They actually are faring worse. And so, the "it gets better has to be," again... It's important but it has to be located in space, and we have to keep thinking structurally about how to address those issues.

1:29:01: Alright. I'm a political philosopher and I work on race, social justice, and inequality. I had two questions, but David anticipated one. I've had the good fortune to teach and do research in really red and really blue states, including at Kansas, where I was a professor of philosophy and law, and Texas A&M. And I was thinking about the discourse of responsibility, 'cause that's typically the first move when one talks about disparities in whatever domain, you know: What role does individual choice play? I really appreciate the answer we got on that.

1:29:39: Let me the raise the second issue, which is about democracy. I think, in at least the two presentations I heard, one way to think about addressing disparities is thinking about change from the top, the lawmakers, the policy-makers, or even engaging with sort of a broader public of voters. I'm curious to know if you all can expand on what role education might have for the people who are disproportionately impacted by the kinds of disparities you're calling attention to? How might the kind of research you do empower them for civic engagement that is aimed at transforming some of the structures that you all are sort of identifying as significant in accounting for disparities? I mean, what kind of moves can we make to show that this research can arm individuals for a certain kinda democratic engagement to alleviate some of these issues? Not to say that the top-down approach isn't important; we all know it is, but that's just one dimension.

1:30:58: One of the things that's done in the SAFE Lab is to create a digital scholars lab where we work with young people, in this example, from Brooklyn, to educate them on the things that we're learning about their social media and digital footprint. In the research, we were learning that there were all these consequences that were associated with a digital footprint, but these young people had no idea of those consequences or what could happen to them once someone got ahold of their social media communication. And so, we wanted to present that research to them, and then also equip them with tools to be able to create their own solutions. And so, this past summer, we brought four young people from Crown Heights and gave them some didactic training sessions of what we were learning, and then also gave them some research tools to come up with some of their own solutions. And these were young people that were beginning to think about college; some were enrolled in their first year. We had one young woman that came and said, "Well, I wanna be a beautician." And then after the end of the program, she said, "Actually, I wanna be an engineer because now I realize that I can do something about some of the challenges that are happening in my community through technology."

1:32:09: One of the things we've always done in my labs is to actually hire young people from the community to be advisory board members or... But we actually hire them. We don't treat them as participants, we don't just fete them and say, "Thanks for your time." And in every single project, when we debrief, when the project is over and we keep in contact with the young men and young women, has been that having the University of Michigan in their résumé now as a paid employment opportunity completely changes their trajectory at work, 'cause they were vetted by such a strong institution in the community, they were went through background checks, they have glowing letters from several of us.

1:32:48: It really changes access to both education and income attainment in their own lives, separate from research. One of my good mentors, who's in the audience, once had a talk with me on a Saturday afternoon when I was a grad student here, and she said, "You know, you love teaching." And I said, "Yeah, I do." And I think that's actually the biggest structural intervention I will ever do. I can build a program with 300 participants and hope that it works and then ask Population Health to scale it up, or I can train master's students, 100 each year, in how to do their work in their communities, and how to do it better. And when you start then adding that up, and they become teachers in their own spaces, it actually is really powerful. I don't wanna lessen the fact that Michigan creates a space where we pay it forward.

1:33:43: I just wanna give you examples of two studies that show the promise in this area. One is the Truth campaign. The Truth campaign was a campaign about the negative effects of tobacco. It got money from the tobacco settlement, but unlike traditional anti-smoking campaigns, the Truth campaign was directed by young people, targeted at young people. And they were informed with the ways in which the tobacco industry had worked to pull the wool over the eyes of young people in terms of the kinds of advertising they had done. The Truth campaign was less about the negative effects of tobacco on your health, but more about the way the tobacco industry had exploited people and treated them as if they were stupid. And what the research on that campaign found: It was more effective than most campaigns, and it had a disproportionate impact on African-Americans and Latinos. It's an example of empowering youth to... With knowledge, and with information, and they're using that knowledge... Has a big effect.

1:34:56: Other, second quick example, Lalonde and colleagues in Canada did a study of First Nations communities, the indigenous population of Canada. That group had the highest levels of adolescent suicide, at the time, of any group in the world. And when they looked at over 1,000 different communities, found that although the group as a whole has the highest level of adolescent suicide in the world, more than half of the communities had had no suicide in the last five years. They tried to identify how could they distinguish those communities that had low rates of suicide versus those that were high? And they came up with this scale that measured, one, the extent to which that community had been involved in fighting for their rights with the Canadian government, and fighting for control over their resources and the services delivered to their communities, one, that was one.

1:35:54: They had like five indicators of that. And the extent to which those communities had a place in their community where they celebrated their culture and taught the next generation their culture. And there was a dose-response relationship between the number of those things the community had and the risk of lung cancer... Of suicide, sorry. And for those communities that had all of those factors, there was no adolescent suicide. It wasn't an intervention, but it was kind of identifying what's naturally occurring. Those people who were well-informed, and who were fighting for their rights, and looking to the future, and training the next generation, seemed to be doing much better on an important health outcome.

[background conversation]

1:36:44: Sorry about that, everyone. I was so engrossed in the presentation and the answers to the questions, I lost track of time. But thank you to all our presenters. That was wonderful information. And thank you for all for coming.


1:37:04: Yeah, great panel. Let me thank Jorge Delva, who organized the panel, put together a great group.