Press "Enter" to skip to content

Op-Ed: Maryland Discipline Study Shows Usual – But Misunderstood – Effects of Policies on Measures of Racial Disparity

Opinion Section: Op-Ed
Agree? Disagree?   Respond to and we’ll publish your point of view on this issue. Or look here for details about how you can start your own Readers Speak! story topic. The views expressed in all Readers Speak! and other opinion pieces do not necessary reflect the views of The Gunpowder Gazette.  

A September 2019 study titled “Disproportionality in School Discipline:  An assessment in Maryland through 2018” by the Regional Educational Laboratory Mid-Atlantic, an entity funded by the U.S. Department of Education’s Institute of Education Sciences and administered by Mathematic Policy Research, evaluated Maryland’s efforts to reduce racial disproportionality in suspensions and expulsions by making removal from school the last resort for addressing discipline problems.

The following was its key finding regarding patterns of disciplinary removals for the period between the 2009-10 and 2017-18 school years:  “Exclusionary discipline rates declined for all subgroups in Maryland over the past decade; however, Black students and students with disabilities continued to be suspended and expelled from school at more than twice the rates of other students.”

The finding, while technically correct, was misleading in certain respects.  First, “more than twice” understated the differences at the end of the period.  The study showed that for the 2017-18 school year, the black rate was actually 2.9 times the rate for other students (7.6%/2.6%) and the rate for students with disabilities was actually 2.4 times the rate for other students (8.9%/3.7%).  Second, rather than remaining at those levels after little or no decline, as phrasing and context suggest, the differences had increased.

Estimates of the earlier rates then can be derived from the study’s Figure 1 (on page 4) indicate that in the 2009-10 school year the black rate was only 2.6 times the rate of other students (11.0%/4.2%) and the rate of students with disabilities was only 2.1 times the rate of other students (12.5%/6.0%).

Click to Enlarge

Thus, contrary to expectations promoted by both the U.S. Department of Education and the Maryland State Department of Education, reductions in school removals had been accompanied by increases, rather decreases, in ratios of the rates of blacks and students with disabilities to the rates of comparison groups.  That, however, ought not to be a surprise.

As I explained in a August 5, 2013 commentary [Baltimore Sun: The Paradox of Lowering Standards], generally reducing adverse outcomes tends to increase, not reduce, percentage differences between rates of groups that differ in their susceptibility to the outcome.  I illustrated the point with hypothetical test score data showing how lowering a cutoff, while tending to reduce percentage differences between pass rates of higher- and lower-scoring groups, tends to increase percentage differences between the groups’ failure rates.  Pretty much everything works like this. That is, whenever there occurs a change in the frequency of an outcome and its opposite, percentage differences between the increasing outcome tend to decrease, while percentage differences in the corresponding decreasing outcome tends to increase.

An easier to understand illustration of why reducing an outcome tends to increase percentage differences in rates of experiencing the outcome may be found in actual Department of Education data on public school suspensions.  Typically, the ratio of the black rate to the white rate is smaller for one-or-more suspensions than for multiple suspensions.  For example, in Maryland during the 2013-14 school year, the black-white ratio was 3.3 for one-or-more suspensions (9.1%/2.8%) and 4.6 for multiple suspensions (4.1%/0.9%).  So, if all students are given alternative punishments for what would otherwise be their first suspension, the ratio of the black rate of one-or-more suspensions to the white rate of one-or-more suspensions will tend to increase.  On the other hand, those same figures would show that percentage differences in avoiding one-or-more suspensions will tend to decrease.

Some of the most compelling evidence that generally reducing public school suspensions will increase percentage racial differences in discipline rates in the overwhelming majority of cases may be found in 2015 study by the College of Education of the University of Maryland.  The study showed that during a period of substantial reductions in suspensions in Maryland schools between the 2008-09 and 2013-14 school years, that ratio of the black rate to the white rate of one-or-more suspensions increased from 2.3 (11.4%/4.9%) to 3.0 (7.8%/2.6%).

Click to Enlarge

More noteworthy, however, an Appendix to the study (at page 7) showed that in 20 of the 23 Maryland school districts for which data could be analyzed, the black rate had a smaller proportionate decline than the overall rate.  This means that in those 20 districts, the ratio of the black rate to the rate of other students increased.

I attempted to explain this issue to the U.S. Department of Justice in June 2012, and in August 2015 and April 2017 to the U.S. Department of Education, as well as by means of an extensive handout at a meeting with agency officials in March 2018.  I also attempted to explain it to the Maryland State Department of Education by letter in June 2018.  None of the agencies have yet shown any signs of understanding the issue.

Meanwhile, both agencies, including by means of government-funded activities like the September 2019 study, continue to promote the belief that generally reducing disciplinary removals should reduce the ratio of the black removal rate to the white removal rate.  And by doing so they affirmatively promote racial mistrust.  For when measure of racial disparity increase in the face of policies that governments lead the public to believe should reduce those measures, observers who believe that racial bias plays a substantial role in those disparities will commonly believe that racial bias must be increasing.

A similar problem is brewing with regard to the Baltimore Police consent decree.  For the decree is premised on the belief that reducing adverse criminal justice outcomes by means of things like diversion programs will tend to reduce, rather than increase, the measures of racial disparity on with the Department of Justice relied in bringing suit.

But as I recently discussed, [The Federalist Society: Usual, But Wholly Misunderstood, Effects of Policies on Measures of Racial Disparity Now Being Seen in Ferguson and the UK and Soon to Be Seen in Baltimore], we will likely be seeing increases in those measures in the near future, just as we are seeing in other jurisdictions where policies were premised on mistaken understandings of the effects of reducing adverse outcomes on measures of racial disparity.

The Gunpowder Gazette, LLC - P.O. Box 183, Monkton, Maryland 21111 - All Rights Reserved