Uncovering Inequities in Special Education: A Conversation with Dr. Paul Morgan

Dr. Paul Morgan is a researcher investigating disparities in disability identification and treatment during childhood. His research has found that students of color are likely to be under-identified for special education services despite a prevailing narrative that over-identification is always the issue. 

Dr. Morgan discusses the methodology of his studies, the importance of investigating a fuller set of explanatory factors for reading difficulties, and why strong claims about special education being a wholly negative "dumping ground" are not supported by the evidence. 

Dr. Morgan also discusses why there has been resistance to accepting the data, and shares findings from his latest study on the prevalence of reading difficulties. 

Transcript

Lauren Clouser:

Welcome to The LDA Podcast, a series by the Learning Disabilities Association of America. Our podcast is dedicated to exploring topics of interest to educators, individuals with learning disabilities, parents and professionals to work towards our goal of creating a more equitable world. Hi everyone. Welcome to The LDA Podcast. 

I'm here today with Dr. Paul Morgan, who's a researcher and Empire Innovation professor and Social Health Equity Endowed professor at the University of Albany. Dr. Morgan's work investigates disparities in disability and treatment during childhood. Dr. Morgan, we're so excited to have you here with us today.

Dr. Paul Morgan:

It's my pleasure, Lauren. Thank you.

Lauren Clouser:

So, to start us off, would you tell us a little bit about your background?

Dr. Paul Morgan:

Sure. So I was a special education teacher for a number of years outside of Richmond, Virginia. I pursued my PhD at Vanderbilt and have been working as an investigator, social scientist, since about 2004. I was at Penn State for a number of years as a Professor of Education and Demography and led a research center there. And now I'm at the University at Albany, one of the SUNY schools, where I'm a professor, as you noted, and direct an institute that's focused on understanding and addressing social and health inequities locally, regionally, globally, nationally.

Lauren Clouser:

So what led you to look into disparities in special education identification?

Dr. Paul Morgan:

I was working with several other colleagues. We started to look at trying to understand how children were being selected into special education or identified. The data that we were using is a data set collected by the U.S. Department of Education, the National Center for Education Statistics. That particular data was nationally representative and followed kids across elementary school and actually into middle school. And the data had a lot of measures of kids, their own characteristics, things going on in the home, or in the school, and their community. And so it allowed us to examine a number of factors that seem to be associated with, or predictive of, being identified as having a disability. And one of the findings that came out of that is contrary to what had been reported previously.

When we adjusted for other explanatory factors, we observed that children of color were relatively less likely to be identified as having disabilities. And so we presented that in 2008. We published those findings in 2010 and have conducted a series of investigations following that to further replicate the findings.

Lauren Clouser:

And could you tell us a little bit about some of those factors that you had mentioned that weren't necessarily looked at before that can lead to incorrect identification?

Dr. Paul Morgan:

Sure, sure. So to step back, just to give a bit of context. Soon after special education began as a system of supports for students with disabilities, there were anecdotal reports that students of color seemed to be more likely to be identified as having disabilities, and so receiving special education. Both anecdotal reporting and descriptive reporting. And that over time led to some policy changes, legislative changes to the Individuals with Disabilities Education Act. 

And there's been concern in the field for quite some time, very legitimate, that we don't want to identify children having disabilities because of their race, ethnicity, or some other immutable, irrelevant background characteristics, just like we don't want to avoid identifying children who do have disabilities because of these background characteristics. So the reporting and the research, up until fairly recently, has consistently reported that children of color are overrepresented in special education. 

So if you look at the percentage children who are Black or Hispanic in special ed relative to the percentage of Black or Hispanic children in the general education population, for example, or if you do something a little bit more complicated and compare the percentages of Black and Hispanic children in special education relative to general education, and do that same calculation for white children, the percentage of white children in special education relative to general education, and you start to compare these, you get some evidence that children of color are more likely to be receiving special education. And that's led to suggestions and claims that special education is biased in the way that it identifies children as having disabilities.

And again, that's a very legitimate and serious concern that we should examine. The problem with the prior research is that to understand whether there's bias operating in the system, there's typically two ways to test that. One is called differential treatment, and the other is called differential impact. So if you look at differential treatment, the standard is to look at children who seem otherwise similarly situated, particularly in the characteristics that relate to the condition. 

So if I was interested in understanding disability, say, learning disabilities, I would try and compare children who seem similar in terms of their academic achievement, which would be kind of one way to test it. Or if I'm looking at kids in terms of diagnosis for ADHD, I would look at children who seem similar in terms of their attention and behavioral regulation. And sometimes you can extend this to other conditions. Asthma, for example, or many other health conditions. But the standard in differential treatment would be looking at kids who seem similar in terms of the characteristics that would predispose them to the condition, and then see whether some groups based on background characteristics are more or less likely to be identified. And this standard has been applied to other issues too, so policing, suspension, many other conditions. And so the prior research in special education didn't approximate those kinds of contrasts.

 So there was really no research that looked at kids who seem similar in terms of the factors most relevant to being identified as having disabilities, and then see if there's some evidence of differential treatment by, say, race, ethnicity, or some other background characteristic. So we attempted to approximate those kinds of contrasts. And so we controlled for academic achievement or behavioral functioning. And we also controlled for other factors that might confound the estimates for race or ethnicity, or otherwise explain initially observed associations between race and ethnicity. So we took into account families' SES and other socioeconomic status and other factors. And so in doing that, what we observed repeatedly is children of color seemed to be less likely to identify as having disabilities in analyses accounting for and adjusting for these other explanatory factors.

And we initially observed that in 2008, we reported on that in 2010, and have somewhere around 20 additional studies that have replicated that finding. And that finding has also been replicated by other researchers.

Lauren Clouser:

We're seeing some cutbacks on federally collected data and education. So would you be able to tell us how that's going to impact studies like this?

Dr. Paul Morgan:

Oh, it'll definitely impact studies like this. So as one simple example, the data that we've relied on beginning in 2008, as I mentioned, that study was called the Early Childhood Longitudinal Study Kindergarten Class 1998, 1999. And that data was collected by the U.S. Department of Education. It followed children from kindergarten to eighth grade and included many assessments of their development. The Department of Education then did another cohort, Early Childhood Longitudinal Study Kindergarten Class 2010, 2011, that followed kids from kindergarten to fifth grade, and included many measures of the children's development. Again, that data collection for the second cohort stopped in 2016.

The department was collecting new data, Early Childhood Longitudinal Study Kindergarten Class, what was called the 2024 cohort, and was going to follow kids again from kindergarten to fifth grade. A new sample, a more recently assessed sample of children with extensive measures of children's development. That would have been the third ECLSK study. And it would have allowed researchers, including ourselves, to further examine these and additional issues around children's development, and special education disability identification. But that third cohort was canceled just outright by the recent DOGE actions. So what's the consequence of that? 

We won't know how these kinds of disparities are occurring with more recently collected data. As a result, the department's been cutting back on NAEP assessments, including long term trends that we would be interested in examining. So, you know, from my standpoint, as a social behavioral scientist, I'm not in favor of these actions. From my perspective, it's shooting ourselves in the foot in terms of knowledge about how to help children early on in terms of development and supporting them with high quality data and research. So these actions are counterproductive.

Lauren Clouser:

So I wanted to ask too. We definitely know the consequences of students not receiving the special education education services that they need. Are there consequences to students being over identified for special education?

Dr. Paul Morgan:

There is some research, a small number of studies that are suggestive of, for children on the margin, children who are identified but may not have disabilities. It's sort of the inference from the research that the identification can or is associated with or predictive of some evidence to suggest that it may be counterproductive or harmful. Why might that be? Lower expectations from teachers, kind of a lower feeling of well-being by the students. It's not a very large literature base. Certainly we need more studies of this. Of course there's contrasting literature that's suggestive that disability conditions can adversely affect kid's performance in schools or their healthy development. Untreated disabilities, I don't think the prognosis is good. Many kids aren't on the margin or close to the fuzzy boundary condition of whether they have a disability or not.

Many other kids have severe difficulties that require specialized long term types of supports. So if I was going to summarize it, I would say children who have disabilities generally benefit from special education services. There are multiple studies that I would say are well conducted, quasi-experimental studies that find that the receipt of special education services for children with disabilities is generally positive, including on their academic achievement. There's some evidence that these services have long term benefits to children in terms of later educational attainment. So you know, in net, I think there's a good case to be made that special education services are generally beneficial, except maybe for kids on the margin where it's a little fuzzy or unclear about whether they have disabilities or not.

Lauren Clouser:

So there hasn't been as much research finding that students of color can also be underidentified for special education. But now you've published several...you said about 20 studies on this topic. Why do you think that there's a bit of a slow uptake on accepting this data?

Dr. Paul Morgan:

That’s a good question. I would say that there's been a bit of a narrative about our field. Every field can be improved. We're talking about services being provided in schools. Any time we're providing services or trying to do something right, there can be ways to make improvements. We can improve dental care, in a sense. We can improve lots of sorts of government services.

So certainly there's room for improvement. I acknowledge that. But I think there's been a bit of a narrative that special education is a dumping ground, it's a way to purposely segregate children, that it's unambiguously harmful. And I just don't think the research supports those claims. I think there's been a lot of strong claims based on what I would characterize as often weak evidence. So I think that our field has been, I would say, maligned. Now, as I was referring to before, there was descriptive reporting, anecdotal descriptive reporting about overrepresentation.

And we find that ourselves. Descriptively, you'll see evidence that the percentages of children of color in special ed seem to be greater than the percentage of children who are white in some relativistic way. So not disputing that, we have reported that ourselves. The question is, is that explained by bias in how schools are identifying children as having disabilities? And there's been a mismatch in the claims versus the evidence for the field. So attempting to approximate these otherwise similar contrasts, those studies are relatively rare, relatively new. I would say as an aside, they're not perfect. And there are limitations to our approach.

So, for example, we can't directly observe whether the children have disabilities or not. So there's certainly limitations for our approach. But I think generally our studies allow for a fair test of this potential bias explanation.

I would say there's lots of reporting in other fields, including in public health, of potential bias. But it tends to run in the other direction that we've often seen in education. Around special education and in public health, the bias seems to run in a direction that children of color are less likely to be appropriately identified and receive health services. So our results are very consistent with those findings. And I can go to that more if you'd like. There are other researchers that have replicated our findings. There is more work to be done. I think replicating our findings with other data is certainly a good idea.

But I think there has been resistance in part because this narrative began based on, I think, relatively weak evidence. So it's kind of baked-in a bit. By some, there's an inclination to view special education negatively, which I think is unfortunate. And so I can go into a number of the reasons why I think this is occurring. But, you know, as scientists, what we're supposed to be doing is investigate these issues objectively, partially, agnostically. If we had found over identification, we would have reported it.

Our methods allowed for either over or under identification. But what we've in fact observed repeatedly is evidence of under identification. I think what we need to do as a field is, approach these issues carefully, conduct careful studies to try to understand how special education is working, how this election process, who's receiving these services. I think investigating that carefully, systematically, welcomes other attempts to do similar studies. And it may be that another study, other types of findings emerge in other contexts or using other types of methods or other kinds of data. That's how science works. 

But I think one thing that I would encourage practitioners, policymakers, and researchers to consider is wholesale claims about special education as only a negative, only a dumping ground. I think those are not warranted or particularly credible, to be honest. So I would encourage viewing services as potentially beneficial.

But of course, we want to make sure that we're doing all we can to improve the quality of services being provided. We want to be careful about the possibility of bias, and there may be bias occurring in certain cases or certain contexts. That's important to know and address. But I think strong claims about special education as a dumping ground, special education as wholly negative, I just don't think those are particularly well-supported by the best available empirical research. And so I don't view them as credible.

Lauren Clouser:

LDA has a core principle focusing on disproportionality in special education, and it does focus on under identification because in practice, our members recognize this issue, even though the prevailing narrative was over identification.

Dr. Paul Morgan:

Yeah. So I will just take a moment to applaud LDA, because I've seen you as the only learning disabilities association that's tried to update its position statements based on the most recent research around racial and ethnic disparities in disability identification. And by and large, what I've seen from the other organizations is an insistence, which I think is misaligned with a good number of studies at this point, that significant disproportionality, initially observed racial differences in special education are because of bias. And so the assumption is that over identification is occurring. 

And from my standpoint, LDA has updated its position and policy statements to reflect the most currently available, and what I would characterize as the best available, research on this topic. And I do think that they should be commended for that. And I do think that it's right that there seems to be evidence that a number of children are not being appropriately identified as having disabilities, and our analyses suggest that's attributable to race and ethnicity. And like I said earlier, that's not what we want to be seeing.

We don't want to identify kids as having disabilities who don't have disabilities because of their race or ethnicity. But at the same time we're obligated, as a field. And our studies find again and again that children are more likely to be identified if they're white and from English-speaking households. And that is a type of inequity that we should be trying to address and ensure that our identification procedures and our eligibility evaluations are striking the right balance and not somehow missing kids who do have disabilities because of these background characteristics.

Lauren Clouser:

Yeah, and I wanted to ask too about your latest study, which is not yet published, but I wanted to give you a chance to talk about it. The findings were that Black, Hispanic and Native American students are especially likely to experience reading difficulties in elementary school. So I just wanted to talk to you a little bit about that study and also ask if underidentification of learning disabilities could play a role in those risks.

Dr. Paul Morgan:

Yeah, it certainly, the short answer is it certainly could. So I guess maybe the connections between the two, our prior work and the new study, is when we've done our analyses, the strongest explanatory factor, the strongest factor associated with or predictive of disability identification, particularly learning disability identification, is their academic achievement. The student level academic achievement is a very strong explanatory factor of whether children are being identified as having disabilities, which is not surprising. It makes sense. You're supposed to be identified as having a disability and receive special education services because you have a disability that's adversely affecting your educational performance. And academic achievement would be sort of a main indicator of academic performance. So as you alluded, I think we've experienced what might charitably be called pushback for studies that we've done. And I don't think that others have maybe understood academic achievement seems to be related to disability identification.

And schools, I think, from my standpoint, from the data that we've seen, are likely to select for evaluation identification kids who seem to be struggling academically and behaviorally. And so that's sort of the group that schools are focusing on in terms of provision of special education services, just like they're focusing on for gifted services children who are displaying high levels of academic achievement. That's sort of the group that gets selected for gifted ed, high achievers. Okay, so children who are struggling academically are especially likely to be identified for special ed. And I don't think the field has really understood very well what sort of composition of children are displaying academic struggles or academic difficulties. So in our new study, we examined who's likely to be displaying reading difficulties as operationalizes performance in the lowest 10%. Students for our study who have reading difficulties are displaying readiness even in the lowest 10%, below the lowest 10th percentile on a well designed test of reading and treatment.

And what we find in the new study is there are large racial ethnic gaps in the relative percentages of children displaying reading difficulties at kindergarten entry. So it's something like about 5 to 6% in children who are white or Asian versus about 15 to 20% of children who are Black, Hispanic or American Indian. And so those differences are quite large at kindergarten. Just as an aside, you know, most kids, white, Black, Hispanic, American Indian, Asian, are not displaying reading difficulties. The large percentage of children across racial and ethnic lines are not displaying low levels of reading achievement as we've operationalized here. But when we focus on relative percentages, greater percentages of children who are of color are displaying reading difficulties at kindergarten relative to children who are white, for example. And those differences are large across elementary school.

And in fact, the percentage of children who are Black who just experience reading difficulties actually increases across elementary school. So it starts around 14% and ends about 20% in fifth grade. And we also find that for repeated reading difficulties, so severe and persistent reading difficulties, so kids who are displaying reading difficulties not just at one particular grade, but across several grades, which, you know, those kids could reasonably be characterized as with or at risk for reading disabilities or dyslexia. The final step of our analysis is to examine, why is that, trying to understand what other factors might explain these initially observed racial and ethnic differences. And the factors that are very strong consistent explanatory factors are the family's SES and the children's early academic skills in reading, but also in math and to some extent in science and their executive functioning, particularly their inhibitory control and working memory. So those sort of small set of factors, family SES, initial academic achievement, and initial executive functioning, largely to fully explain the initially observed racial and ethnic differences in the risk for reading difficulties across elementary school. In fact, once we account for these other explanatory factors in our study, there are others. The initially observed risks attributable to race and ethnicity are fully explained.

So you know the odds ratios are going something like 2.2.5 to less than 1 for children of color relative to white children. So in other words, the initially observed risk attributable to race and ethnicity is fully explained by these other explanatory factors, suggestive of to us that these factors are the ones that we want to intervene upon through economic and educational policies. And if we could fully ameliorate or address and intervene upon these factors, one could reasonably conjecture that the racial and ethnic differences initially observed would no longer occur. So if we want to help children from experiencing reading difficulties across elementary school, including children of color, then targeting these other modifiable factors seem, from our standpoint, to be a promising direction, a good bet. So that's our new study. Let me just say two more things, if I may. One of the conclusions from our study is because children of color are more likely, as initially observed, to be at greater risk for reading difficulties.

An implication of that is if we're providing early intervention services to children who are having reading difficulties, then we shouldn't be particularly surprised that there seems to be racial and ethnic differences in terms of who's receiving those interventions. The second thing is the field around reading difficulties, disabilities, dyslexia. There's been calls for a more expansive set of investigations, and the concern has been the focus on a single or a small set of cognitive-type skills. Trying to find a cognitive marker for reading disability may not be particularly fruitful. So it doesn't seem to be the case that reading disability is the result of one particular cognitive skill, say, phonemic awareness, rapid, automatic naming. And there's been calls for investigations of reading disabilities and dyslexia, that take into account a fuller set of explanatory factors, including across different contexts.

So these are sometimes called multivariable risk and resilience models. So you're looking at factors across not only student cognitive or academic skills, but other types of skills, executive functioning, looking at characteristics of their homes or their schools. And so there have been calls for research that examine the occurrence of reading disability or dyslexia using a much richer set of explanatory factors across multiple contexts. Kind of a biopsychosocial type of modeling, and our study helps provide that. So we use a well validated theoretical framework called the antecedent opportunity propensity framework to examine a very extensive set of explanatory factors as well as severe and persistent reading difficulties across elementary school. And we identify a set of factors that seem to be related to those risks, including those that are modifiable through economic and educational policies. So the reason I'm mentioning this is I think the results of our study are useful, not only trying to understand the kind of racial and ethnic differences and the risks for reading difficulties, but it also could be used to inform theoretical and applied work specific to reading disabilities and dyslexia because of the extensive nature of the explanatory factors that we investigate.

Lauren Clouser:

Yeah, I'm really glad you brought that up. That could be its own podcast episode easily. Okay, well, you're giving me ideas.

Dr. Paul Morgan:

Okay. All right, I appreciate it. Happy to have a follow up conversation about it in any way that's helpful.

Lauren Clouser:

Oh, yeah, that would be fantastic. But Dr. Morgan, before I let you go, I just wanted to ask, is there anything that I didn't ask that maybe you wanted to talk about? Anything from your research to any practice recommendations, anything that you wanted to maybe end on?

Dr. Paul Morgan:

I think that, you know, like any field, there's ways to improve our field. But our field has a lot to offer in part because I think we put high standards on ourselves in terms of the scientific rigor that we try and use where we come from. Not all special education researchers maybe would agree with this. But there's traditionally been an emphasis in special education research on what might be called positivistic, you know, quote unquote, objective investigations that are sort of empirical in nature, and those could be carefully done as quantitative as well as carefully done qualitative work. I think in some sense there's been a debate maybe, or even a battle within our field about whether those traditions are valid. There's been a lot of problematizing about sort of empirical approaches, but I do think those approaches have a lot to offer. And I think that our field has made a lot of advances through use of those kinds of methods. So I think we have a really good understanding of how to teach reading that may have not come about using other types of methods.

I think we have a good understanding about how to teach math effectively to students that may have disabilities. I think we have a lot of good knowledge about how to manage classroom behaviors in a way that's likely to be effective. And so I think there's a lot to commend our field. Not that we got everything right, but there's a lot of great research that's come out of our field that could be used to help many kids. And I hope that there's a continuing, continuing emphasis on sort of empirical approaches to investigation that we continue to uphold. And I think in some ways it reflects the best of our field in terms of what knowledge we've generated and what we have to offer. I think more broadly, there's a lot of kids beyond those with disabilities that could be helped with the kind of approaches that we've developed in special education. How to teach children academically, as well as teach social and behavioral skills.

Lauren Clouser:

Absolutely.

Dr. Paul Morgan:

Keep up the great work that you're doing. It's such an important group of children to advocate for. We have a lot of well-validated information from our field to help children with disabilities, including learning disabilities. And I think those methods could also be used to help many other children as well.

Lauren Clouser:

Thank you for listening to the LDA podcast. To learn more about LDA and to get valuable resources and support, visit LDAAmerica.org.