Stanford partners with San Francisco schools to get to the bottom of racial disparities in special education
For decades, educators and researchers have raised concerns about students of color, particularly Black students, being disproportionately enrolled in special education. At San Francisco’s public schools, disparities persist, despite the district’s many efforts to address the problem.
A new Stanford research initiative is reaching further upstream to find solutions. Led by Stanford education Professor Alfredo J. Artiles, a team of interdisciplinary researchers from across the country has partnered with leaders at San Francisco Unified School District (SFUSD) and California Education Partners to identify what’s driving the disproportionate enrollment and, based on their findings, design new approaches to change those conditions and create an environment where Black students can thrive.
“Some kids need special education, but when you look at the data on how students in different subgroups are identified with a disability, you see trends that make you question how and for what purposes the tools of disability policy are being used,” said Artiles, the Lee L. Jacks Professor at Stanford Graduate School of Education (GSE), who is leading the initiative as part of a research-practice partnership between Stanford and SFUSD that began in 2009. “There might be racialized patterns that make you wonder if disability identification is being used to push students out. How can we use special education to provide the resources that some kids need, without making it a tool for exclusion? That’s the challenge.”
Here, Artiles talks more about the persistent issue of racial disparities in special education, what makes this new initiative different from previous efforts, and what they’ve found so far.
GSE Professor Alfredo J. Artiles
What has past research shown about disproportionality in special ed?
The first time this was addressed in the research community was in 1968, in an article published in the journal Exceptional Children, and from that point on, there have been debates about this issue.
Disproportionality isn’t just about overidentification — some groups tend to be underidentified for intellectual, learning, and emotional disabilities, and that’s a problem, too. But most of the attention has been on patterns of overidentification.
Two explanations have prevailed over time. One view attributes disparities to poverty – that children who are disadvantaged by high poverty don’t develop the dispositions and skills that prepare them to do well in school, so they struggle. We know that poverty, particularly persistent and intergenerational poverty, has a significant impact on children’s development. But Latinx groups, for instance, tend not to be overidentified for special education at the national level, even though there are high poverty levels in this community. There are Asian American subgroups that experience high poverty levels but do not experience disproportionality in special ed.
An alternative explanation focuses on factors in schools, like teacher bias or a school climate that is less conducive to learning. In other words, it is not clear if these students’ difficulties in school are because of a lack of opportunity or a low-quality education. We have less evidence on the role of these types of institutional and systemic factors in producing disproportionality, and the findings are somewhat mixed.
But nobody is questioning the overrepresentation of Black students, right?
Some researchers do question whether overidentification exists, finding that students of color are underidentified and reporting an absence of racial bias. But scholars have raised a number of conceptual and methodological concerns about these studies.
Another issue is that the federal government started requiring districts to monitor special education placement by race in the late 1990s. If districts report disproportionate levels of identification for distinct groups, they are cited and have to develop and implement plans to address the problem. Federal guidelines to determine disproportionality were also vague. So districts began to game the system by changing the criteria required to determine disproportionality, often resulting in a relatively small group of school districts being designated as disproportionate.
You’re now working with SFUSD, which has documented disparities in special ed in the district and tried for years to address them. What’s different about the approach you’re taking?
For one, we recruited a group of scholars from a variety of disciplines — sociology, learning science, linguistic anthropology, data visualization, digital technology – many different fields in addition to special ed, because we wanted to disrupt the way the issue has traditionally been framed. The district has worked hard over the years to deploy policies and interventions to address the problem, and we wanted to find a new way of understanding it.
We’re focusing our attention on the conditions that exist before a student is identified with a disability. We spent the first year reviewing and analyzing district-wide data on the history of the problem and interventions to address it. We consulted administrators, educators, parents, and community and religious leaders about the problem and alternative solutions.
We’re also prioritizing the design of a system in which Black families are part of ongoing decision-making in the district. Throughout SFUSD’s history of disproportionality, Black parents have provided a lot of input and suggestions. But they’re not experiencing meaningful participation, and many express involvement fatigue, like, ‘Here comes another team of researchers to do focus groups, and then they go away and nothing happens.’ A shared governance model offers a way of participating in a legitimate fashion about investments, priorities, problem-solving, and so on at the district level, where systemwide change happens.
“We found that on average, Black students at the elementary and K-8 schools were five times more likely than all other students to be referred to special education. At 50 of 70 of the schools in the district, Black students were at least three times more likely to be referred, which means the problem isn’t driven by a few schools — it’s a district-wide problem.”
What’s the reason for looking at students’ experience before they’ve been assessed and formally identified with a disability?
Prior research on this problem has looked at identification rates – how many students in various groups are being placed into special education and the predictors of identification. We wanted to examine the precursors of these processes, to see who was getting referred.
The referral decision is a key marker that a student is already on a different trajectory. By the time you get to that point, the system has likely done interventions, provided extra support and accommodations, and you have a high chance of being identified or designated as having a disability. We’re interested in understanding the support systems available and how they’re implemented in general education before an educator decides to refer a child for special education assessment.
What has your research revealed about referrals at the district?
We found that on average, Black students at the elementary and K-8 schools were five times more likely than all other students to be referred to special education. At 50 of 70 of the schools in the district, Black students were at least three times more likely to be referred, which means the problem isn’t driven by a few schools — it’s a district-wide problem.
We calculated a referral risk index by school, so we could identify high- and low-referring school sites. Then we began taking a more in-depth look at both types of schools and asking contextual questions: What do they do when a kid begins to struggle? What are the systems of support? Do kids have access to specialists? Are parents involved? We also have access to school climate data collected by the State of California, so we can see the connection between the sense of belonging and trust in relation to the risks of referral.
What will you do with the information?
Right now we’re mapping the distribution of disproportionality, finding the hot spots with high referral rates and exploring factors shaping disproportionality at the district level as well as in a small set of pilot school sites. We’re also broadening our analysis beyond the schools, looking at ‘ecologies of learning’ — features of neighborhoods that are potential resources for learning and nurturing, like libraries, community centers, advocacy organizations, anything that might offer opportunities for kids to stay engaged. They might have nothing to do with disproportionality, but they might allow us to identify assets that could be brought to a school’s efforts. A geographer and a data visualization specialist are collaborating with us to create visualizations of this information.
The data we’re collecting is incredibly rich, and having tools that show the cultural and spatial dimensions of the problem can help us to see it in a new light. The goal is to design district-wide systems that can be sustained over time, and to create learning environments where educators don’t have to resort to expulsion or pushing students out.
This initiative has been supported by a vision grant from the Spencer Foundation, as well as funding provided through the Stanford-SFUSD Partnership.
Faculty mentioned in this article: Alfredo Artiles
