U.S. Department of Justice seal

Will Trump Have the First Numerate Administration?

James Scanlan January 04, 2017

On November 28, 2016, the Department of Justice (DOJ) filed an amicus curiae brief urging the Supreme Court not to grant a writ of certiorari for review of the Fifth Circuit decision in Veasey v. Abbottstriking down a Texas voter ID law.  Given the likely views of the Trump administration on voter ID laws, if the Court grants review, DOJ may well file a brief in support of the petitioner.  Such brief would present the matter rather differently from the agency’s recent filing.

Meanwhile, groups are pressuring the City of Baltimore and DOJ to enter a consent decree addressing racial differences in criminal justice outcomes before the change in administrations.  Even if the decree is entered before that change, however, new leadership at DOJ may modify the decree or enforce it in a manner materially different from that contemplated by the DOJ attorneys who negotiated it.

These situation are among many where civil rights law enforcement actions or policies of the current administration may be substantially modified by the new administration.  The question is whether in effecting those modifications the new administration will recognize the extent to which existing policies have been based on understandings of statistics that are the exact opposite of reality. 

The recent DOJ brief emphasized both the difficulties in securing an acceptable ID and the large percentage differences between rates at which whites and racial/ethnic minorities failed to secure one.  But, as I explained here in early October, the greater the difficulty in securing an ID, the smaller (not larger) will tend to be percentage racial/ethnic differences in failing to do so.    

In the same post I explained that the DOJ’s actions regarding Baltimore police practices were based on the mistaken belief that generally reducing adverse criminal justice outcomes would tend to reduce the proportion African Americans make up of persons experiencing the outcomes.  Such was also the belief underlying DOJ actions regarding Ferguson, Missouri.  But generally reducing such outcomes will tend to increase, not decrease, the proportion African Americans make up of persons experiencing them.

DOJ actions regarding the racial impact of voter ID laws and police practices are merely examples of widespread federal civil rights enforcement policies based on the manifestly incorrect belief that relaxing standards or otherwise reducing the frequency of adverse outcomes will tend to decrease (a) percentage racial/ethnic differences in rates of experiencing the outcomes and (b) the proportions more susceptible groups make up of persons experiencing them.  As noted, such actions actually tend to increase (a) and (b).  Both with regard to mortgage lending and school discipline, this failure of understanding on the part the government has led to the perverse situation where the more an entity covered by civil rights laws accedes to government encouragements to relax standards, the greater the chances that the government (or others) will sue the entity for discrimination. 

But this particular failure of understanding on the part of federal civil rights law enforcement agencies – which, it warrants note, is largely shared by Congress, the courts, and the social and medical science research communities – is but one manifestation of a near universal failure to understand the ways standard measures of differences between outcome rates tend to be systematically affected by the frequency of an outcome.  In consequence of that larger failure, most analyses of racial/ethnic and other demographic differences in outcome rates have been a waste of resources even when they have not been patently misleading.  And those resources have been substantial, as suggested, for example, by the fact that the budget of the National Institutes of Health commonly devotes about $2.7 billion a year to health and healthcare disparities research.

A fuller description of the failings in most analyses of demographic differences may be found in my November 14, 2016 comments submitted to the Commission on Evidence-Based Policymaking.  

Some part of problem lies in the fact that most people find the issues difficult to understand. Indeed, very few people understand even that lowering a test cutoff, while tending to reduce percentage difference between the pass rates of higher- and lower-scoring groups, tends to increase percentage differences between the groups’ failure rates.  Still fewer understanding that improving health care, while tending to reduce percentage differences in rates of receiving appropriate care, tends to increase percentage differences in rates of failure to receive such care. 

But there are also entrenched interests in continuing to fail to understand the issues.  This is especially the case among federal agencies enforcing civil rights law where understanding the issues would entail acknowledging that many longstanding policies have been based on statistical understandings that are 180 degrees off the mark.          

Under new leadership those same agencies will have an interest in highlighting misunderstandings that have underlain current policies, especially if leadership decides to challenge the disparate impact doctrine that has played so large a role in civil rights enforcement policies.  See my May 6, 2016 post titled “Is the Disparate Impact Doctrine Unconstitutionally Vague?” (possibly easier to read at the PDF available here).  Further, an administration that wants to cut waste, and one that simply wants to prove that it is smarter than prior administrations, should have a strong interest coming to understand the full scope of the unsoundness of existing policies and research. 

The length of the November 14 CEBP comments (just over 25,000 words) should make evident that much of the drafting occurred before November 8.   Prior to that date, the comments might well have seemed a fool’s errand, like many similar communications I have sent to government agencies and varied other entities over the years.  The largely unanticipated events of November 8, however, may substantially affect the readiness of governmental entities to rethink the analyses of demographic differences with regard to both civil rights law enforcement and research into demographic differences.  The CEBP comments, I suggest, provide a fair guide for doing that.

Questions? Reactions? Email with your response to continue the conversation. Select comments may be published on the blog.