When Frances Haugen alleged that Facebook, now known as Meta, knew its social media platforms were harmful to the mental health of adolescents, she renewed attention to two key trends within the U.S. The first is the trend toward near ubiquitous use of mobile devices and social media among adolescents. The second is the rise in struggles with mental health among adolescents, including rising rates of major depression and suicidality. A substantially large body of research has accumulated over the past two decades that seeks to identify mechanisms linking the two trends. How large? When several studies investigate similar outcomes, researchers conduct systematic reviews and meta-analyses to summarize overall trends in findings. The number of studies conducted on this topic is so large that we are now seeing a new tier of summaries called umbrella reviews, which are summaries of systematic reviews and meta-analyses. In other words, we have numerous summaries of these research studies, and then we have summaries of these summaries because they are so numerous.

Despite what seems at times like an oversaturation of studies asking similar questions, there remains a critical gap: few studies examine adolescents from communities of color and of those, even fewer study Latino adolescents.

The gap in research is surprising for several reasons. Latino adolescents have a higher rate of social media use than their White counterparts. Moreover, they face greater risks of experiencing adverse mental health outcomes and face steeper barriers to accessing mental health care, including the stigma of being labeled a mentally ill person and potential concerns about how their own legal status or that of their guardians may impact receiving care.

Over the past three years, I have made a concerted effort to address this gap and I plan to build on this work when I join Michigan State University in January 2023, in the Department of Media and Information. My work has been funded by a wide range of agencies and foundations, including one run by Meta, which shows how eager people are to hear the voices of these adolescents. Indeed, this work has uncovered the uniqueness of Latino adolescents and how useful their perspectives are to understanding more broadly the nature of the relationship between technology use and mental health among adolescents.

For example, my research team and I conducted interviews with over 40 Latino adolescents about their social media use and mental health. Each interview lasted an hour to an hour and a half. Several questions were geared toward understanding the relationship between mental health and social media use, and the potential implications. The interviewees identified several ways that social media can harm mental health, like providing an additional avenue for bullying. Notably, they also discussed scenarios where the causal arrow went the opposite direction; that is, where they noticed peers who were struggling with mental health turning to social media to access support and cope.

At times, this was described as histrionic and driven solely by a need to attract attention, but they collectively understood a key barrier that Latino adolescents face: sometimes they have no one else to turn to. Mental health is a taboo topic among Latino communities, and so they felt that some of their peers may not feel comfortable going to their own families for help. Some said these individuals should really be going to a therapist or at least a school counselor but noted that this may be difficult without getting marked as “different” by others. They expressed concerns about how families would react and also about the speed at which gossip about their mental health would spread across the school. Those who were struggling with mental health generally would not make a public post that could get read by anyone, but rather would make posts that were targeted toward a narrow set of close friends they could trust. An in-person conversation was seen as more ideal but given that their schedules were often controlled by adults, this was not always an option. Social media was therefore a lifeline for some Latino adolescents who struggle with mental health.

This is an important and often overlooked point in discussions about the relationship between social media and mental health among adolescents. I have participated in several high school events to provide expertise and answer families’ questions about social media and mental health, and too often I come across an assumption that social media only harms. Yes, it can cause harm, but it can also be beneficial, and this research on Latino adolescents helps drive home the point.

Given how common adolescents like themselves share mental health struggles on social media, we wanted to know how best to support them. We asked them who they would turn to if they were concerned about a friend who, based on their social media account, may be struggling with mental health. The logic here is that we wanted to get a picture of the current mental health resources available to them. To do this, we created a card sorting task. Each card represented something a friend of theirs could share that they are struggling with on social media. The content of the cards were items that are either wellbeing concerns – like self-harm – or issues that could escalate into wellbeing concerns, like getting into trouble with authorities like teachers and law enforcement.

We asked the interviewees to separate the items into different piles: those they would tell an adult about and those they would not. We then asked them to discuss general rules guiding their decisions. Participants were asked via open-ended questions what adult they would tell and why, and also which ones they would not tell and why.

Parents and guardians were the most common choice for an adult they would tell. The interviewees recognized the instrumental support that these adults could provide, such as operating as gatekeepers to mental health care. They felt that if the friend needed access to a mental health professional, an adult like a parent or guardian would be the best choice for mediating access.

However, not every interviewee would turn to parents or guardians for help and several gave reasons for why they would avoid telling them. These included concerns that the post would not be taken seriously by the adult.

Others, particularly boys, were concerned that a parent or guardian would get angry and make the situation worse. What matters here is not whether the parents or guardians would actually respond this way, but that Latino adolescents are worried that they would, which then diminishes the likelihood that they would reach out to them in the first place.

My research team and I are grateful we were able to gather such rich data from Latino adolescents. We took several steps to gain access to the Latino community and engender trust, which was particularly important given the possibility that some adolescents or members of their families were undocumented. We collaborated with a non-profit organization, United Community Center, which serves the Latino community in Milwaukee, Wisconsin and regularly mediates access by researchers. Prior to data collection, they helped us convene a focus group of 10 Latino adolescents, which met at the organization, to provide feedback on our study design and assist in making our questions accessible. This step was important for understanding how best to talk about mental health in ways that limits viewing it as a taboo topic and stifling conversation. The focus group also helped us select the topics to be included in the card sorting task. We revised the study based on their feedback and recruited participants from a youth program run by the organization. The interviews were held in a quiet, private room located in the organization, during a time that the adolescent was already scheduled to participate in the youth program. Collaborating with this community organization was integral to our ability to address gaps in existing research by incorporating the experiences of Latino youth.

A focus on working with a specific community does raise questions about how well the findings generalize and depict broader patterns among adolescents. We were able to conduct two national surveys here in the U.S. to evaluate how well these findings from this specific community generalize. In the first survey, we recruited adolescents from across the U.S., creating a sample that is racially diverse. This allows us to examine the degree to which the patterns observed in the interviews are unique to Latino adolescents. It turns out they are quite unique. We gave survey respondents a scenario describing a friend who, based on their social media account, was likely experiencing depression and then asked them who they would turn to in order to help out their friend. The scenario was based on a standard text used in previous surveys, including the General Social Survey, to describe someone experiencing symptoms consistent with depression. We conducted cognitive interviews with adolescents to modify it in ways that aligned with their own expectations for how the scenario would unfold. We found that Latino adolescents were the least likely of any racial group to say they would turn to a parent or guardian. The concerns we found in the interview study, then, appear particularly salient for Latino adolescents across the U.S.

The second survey is of parents and guardians of adolescents in the U.S. We gave them a similar scenario, whereby we asked them to imagine their adolescent child came to them and said that, based on social media, they were concerned about a friend. We asked them how likely they would help. Interestingly, we did not find any ethnoracial variation in the reported willingness to provide help. Indeed, everyone said they would be highly likely to help.

The results of these studies reveal a discrepancy, in which Latino adolescents think their parents and guardians would not help, but the parents and guardians are just as likely as everyone else to say they would help. The next step in this research program is to work with families to understand and address the source of these discrepancies, and so we are currently talking to various community organizations and are planning on prepping research proposals to develop materials to encourage communication among Latino families to address this issue, and better prepare parents and guardians for operating as gatekeepers to mental health care.

This trajectory of research papers would not have been possible without the initial input from the focus group comprised of Latino adolescents that my research team and I held at the United Community Center to gather input on the design of the interview study. A recent review of research on adolescents found that only 1 out of 8 studies bothered to ask adolescents about their input on research into their lives. Gathering their input is critical to ensure researchers identify ways to approach taboo topics like mental health. Moreover, given the speed at which technologies change, it is important to also learn from them about the most recent ways they use technologies.

To address the dearth of research conducted with their perspectives, my collaborators and I have recently founded a youth advisory board, which is like a standing focus group that will provide input on various research projects. The board is currently known as the Milwaukee Youth Wellness Initiative on Technology (MYWIT). We purposely recruited adolescents from communities of color to join as youth advisors. Currently, MYWIT is comprised of 6 youth advisors (one who identifies as Asian, three who identify as Black, and two who identify as Latino), who have committed to participate for 12 months. In exchange for their input, we are providing them with a stipend and several professional development opportunities, including writing an editorial about their thoughts on the link between social media use and mental health. We also have plans and funding to invite professionals from communities of color to speak with them and share their own experiences navigating college and careers.

MYWIT meets virtually to make it easier for the youth advisors to fit meetings into their schedules throughout the year. This of course introduces a barrier, in that some potential youth advisors may steer away from this opportunity because they do not have access to high-speed internet or an internet-enabled device. Thankfully, we have funding to provide these to any youth advisors in need.

MYWIT is currently providing feedback on the design of a new research study that is funded by Meta, in which we are examining how best to design a chatbot that identifies when adolescents experience bullying on social media and deploys social support. We are conducting focus groups with adolescents from across the U.S. who are racially diverse. Targeting bullying is interesting for several reasons. First, bullying can harm the mental health of adolescents, specifically leading to depression and self-harm, and so a chatbot could facilitate accessing care. Second, given the patterns found in the interview study and survey showing that Latino adolescents are trepid about reaching out to parents and guardians, alternative channels of support are needed. Third, there are ethnoracial disparities in the impact of bullying, in that its harmful effects on mental health are more pronounced among minoritized ethnoracial groups. A chatbot could therefore address these disparities in mental health while also helping mitigate inequities in access to mental health care.

Despite the potential, there exists a fourth reason bullying is so interesting, which is that there is a risk of not being able to identify and address these disparities. Other research indicates that members of minoritized ethnoracial groups have a higher threshold for behaviors that they would consider bullying. In other words, given the same behavior, such as calling someone a name on social media, there would likely be variation across ethnoracial lines regarding whether this behavior would be considered bullying. Minoritized ethnoracial groups would be less likely to say this is bullying, because they have a higher threshold for what constitutes bullying. Furthermore, within these groups, boys tend to be more likely to exhibit a higher threshold than girls.

Researchers currently do not have a good understanding for why distinct thresholds exist. Part of it may be due to ethnoracial variation in family socialization and experiences. For example, families within communities of color tend to be a source of bullying for adolescents, such as making negative comments about their physical appearance and body weight. This suggests the distinct thresholds may be due to adolescents from communities of color being more likely to normalize bullying. However, this does not explain why the effects of bullying are more pronounced for these same adolescents, because if bullying was more normalized, then it should follow that the impact of bullying on mental health would be muted. While the underlying process is not clear, the implications for designing a chatbot are, which is that the chatbot should be sensitive to ethnoracial variation in what constitutes bullying.

This speaks to larger issues in the development of artificially intelligent agents, which is the risk of racial bias if the agent – the chatbot in this case – does not consider such variation in the ways people would label a behavior as bullying. Additionally, because boys and girls also have distinct thresholds, the research problem highlights the importance of taking an intersectional approach when examining ethnoracial variation. Over the next few years, I will be leading an interdisciplinary team that takes the findings from the focus groups to begin designing and testing the chatbot. I am particularly excited to join the Department of Media and Information because of the interdisciplinary approach to research that the faculty embrace.

MYWIT will continue to provide feedback throughout the process of designing and testing the chatbot, along with other projects. I plan on recruiting adolescents from communities of color around Michigan to join MYWIT and will rebrand it as the Midwest Youth Wellness Initiative on Technology to retain the same acronym. I am interested in recruiting youth from the Lansing and Detroit areas, as well as from the growing population of migrant communities in the Western part of Michigan. By expanding who is engaged in the research process, this has the potential to make science more transparent and engender trust in science within these communities. The opportunity may also pique their interest in science and facilitate diversifying the pipeline of future scientists. I see several opportunities to continue funding this endeavor from agencies that are interested in supporting these aims, such as the National Science Foundation and the National Institutes of Health.

The examples of studies I have conducted all focus on social media, but my research portfolio examines technologies more broadly. A colleague and I recently published a study in the Journal of Adolescent Health, in which we identified ethnoracial patterns in the use of different telehealth modalities for mental health care among a national sample of adolescents. Like other studies of adolescents during the COVID-19 pandemic, we found a high prevalence of symptoms consistent with depression and anxiety within the sample. Unique to our study is an understanding of the different ways adolescents are accessing mental health care. We found that video chatting with a mental health provider is very popular among adolescents during the COVID-19 pandemic, but Black and Latino adolescents are more likely than their White counterparts to report using text-based chat to communicate with a mental health provider. Privacy concerns appeared to be a key driver of the patterns we observed, in that text may be preferable over a modality like video where other household members could listen to what is being said. In a working paper, we are using the same dataset to show how virtual learning may at times protect the mental health of Black and Latino adolescents, in part because they report sleeping better than when they attend school in person. Both studies are important for understanding how best to use technologies to support the mental health of adolescents from communities of color.

I am thrilled to be joining MSU and am grateful for the resources many have provided to support my research, including the opportunity to use NEXO as a platform to introduce myself to the community.