“The Error of the Third Type” Samuel L. Myers, Jr. (January-February 1995 P&R Issue)

Policy analysts talk of what William Dunn calls “The Error of the Third Type.” Unlike Type I and Type II errors, well known to students of statistics, errors of the third type refer not to rejecting or accepting null hypotheses when they are true or false. These errors refer to finding the right answer to the wrong question.

The debate illuminated in Lawrence Wright’s hopeless essay, “One Drop of Blood,” is an excellent example of the error of the third type. The entire discussion of the utility of racial classification is a misguided and often dangerous exercise in trying to find the right answer to the wrong question.

Take, for example, former Minneapolis Mayor Don Fraser’s attempt to solve the crime problem by refusing to submit to the state racial data on arrests. Minneapolis has the distinction of having the second highest racial disparity in arrest rates among the nation’s cities with populations over 250,000. (Sister city St. Paul has the highest.) The black arrest rate is 24 times that of the white arrest rate. And why? Because presumably, “racial identities may be arbitrarily assigned by those who are required to collect such information or reported inconsistently by affected individuals. . . , “as Mayor Fraser wrote Attorney General Janet Reno, urging her to change the way racial statistics are collected.

When Mayor Fraser instructed his Chief of Police to refuse to supply to state officials figures concerning the race of arrested persons, what could he possibly have had in mind? That many of the persons labeled “black” were really not black at all? That by reclassification to some other category-perhaps mixed race-the racial disparity would disappear?

Unfortunately, the consistent and alarming discriminatory patterns of arrest will not go away if we simply stop collecting statistics by race. Worse, analysts who wish to document these patterns will be forced to abandon these efforts and anti-discrimination litigation will be dealt a fatal blow.

Why, one wonders, has so little been done in the nearly 20 years since passage of the Equal Credit Opportunity Act to force lenders to cease discriminating against blacks and other nonwhites in mortgage loan approvals? Why only now in the past few years have the Atlanta Constitution, Wall Street Journal and New York Times reported definitively on racially disparate loan rejection rates? Quite simply, the answer is that until 1990 lenders were not required to report race on individual loan applications, so that no one knew for sure whether there existed racial disparities in loan approval rates.

In several recent papers, my graduate students at the University of Minnesota and I have been examining racial disparities in loan approval rates in the upper Midwest. We are fortunate to know not just the race of the applicant but also the race of the co-applicant. Does being in a mixed race household afford the applicant an advantage in the marketplace for loans? The answer is a resounding no! We have documented a pervasive pattern of racial discrimination in mortgage lending in the upper Midwest. And we find a definite disadvantage faced by mixed-race or biracial households.

If people think that separating mixed race peoples from the rest of the unjverse of nonwhites will somehow make these persons better off, they are fooling themselves.

Perhaps there lingers in the minds of some the false belief that either by chance or design biracial or multiracial persons are better off financially or economically than the rest of us colored folk. Lawrence Wright quotes ltabai Njeri as saying as much. I don’t know where Njeri gets her numbers, but they do not confirm the Minnesota experience.

David Waithaka, a talented graduate student at the University of Minnesota, with the support of Wendy Treadwell of the University’s data center, helped me examine the 1990 Public Use Micro Sample Census data for Minnesota. Minnesota happens to have one of the highest rates of interracial marriages in the nation. It also happens to have a very small population of African Americans, Asian Americans, and American Indians, who collectively account for less than 3% of all married-couple households. Unlike what Lawrence Wright will have you believe, it is really very easy to glean from existing Census information mixed and bi-racial classifications. We looked at persons who listed themselves as black but who then listed as their ancestry one of many Northern and Western European nations. Not surprisingly, given what we know of the historically large percentages of mixed-race marriages in Minnesota, there were more black married heads of households who identified their ethnic ancestries as European than those who were simply black with either no identified ethnic ancestry or with African or other ancestries. We even calculated the mean household incomes of the “European blacks” and the “Non-European blacks.” And we found that the “European blacks,” or essentially the mixed-race or bi-racial blacks who nevertheless accept the label black, consistently have lower household incomes than the “non-European blacks.” Of course, this is a tricky exercise because the numbers are all very small in this sample-the PUMS sample is a 5% sample of the state’s population, which only has 940,000 married-couple households, of which only 0.9% are black. Still, it is not obviously the case that mixed-race persons are any better off than the rest of us colored folks. If there is a need or desire to prove or disprove this, existing classification schemes exist already for these purposes.

While the issue of racial classification has interest as a theoretical curiosum, as a practical matter the debate is focusing on the wrong question. It matters not whether I have 1/32 Cherokee “blood.” 19/32 black “blood,”
The Censuses of l980 and 1990 asked distinct questions about race, Spanish-origin and ancestry. This permits us to describe the vibrant tessellation resulting from the many cultures, ethnicities and linguistic groups now represented in the United States. We also have the detailed data required by the Voting Rights Act of 1965 to ensure equity in drawing up legislative districts down to the smallest
and who knows how much Irish, French, Spanish, Cuban and perhaps German (note the name “Myers”) blood. What matters is that I am black. My mother and father are black. My sisters and my daughter are black. And regardless of who my daughter eventually marries, her children will also be black. And as black people in America, we are deemed less able, less deserving and less equal than whites. What matters is race. What matters is not being endowed indisputably with white privilege in America. When bi-racials, mixed-race persons, or Asians and Latinos come to this conclusion, as American Indians long ago came to this conclusion, they will see the benefits of joining forces with other minorities and whites of good will to fight racism and racial discrimination. Not by clever-handed racial redesignations or schemes, but by litigation and legislation on the battlefields of employment, housing, lending and education.