Allessandra Bradley-Burns freely acknowledges she is a successful woman. The self-described “innovator and strategist” and “Huffington Post” contributor credits her elite education at Georgetown‘s Schools of Foreign Service and Medicine for paving the way.

But there’s a caveat: Bradley-Burns, an African American, was accepted to Georgetown through an affirmative action program. She didn’t come from a deprived background—far from it. Attending top-flight primary schools, she grew up blissfully upper-middle class in a community comprised of “doctors, lawyers, professors and architects” yet is indignant at any suggestion she didn’t deserve the special privileges extended to her.

Her story is just one confirming how affirmative action is broken. Under the best intentions, President John Kennedy issued the first executive order 51 years ago to ensure federal hiring practices were free of racial bias; later it expanded to include federal contractors, state and local governments and university admissions. Since then, quotas and race-based preferences have been challenged, overturned and upheld in the Supreme Court. Nobody is neutral.

A new book by legal scholars Stuart Taylor and Richard Sander lobs the latest salvo in the ongoing battle. Mismatch: How Affirmative Action Hurts Students it’s Intended to Help, and Why Universities Won’t Admit It shoots a cannon through the sacrosanct assumption that racial preferences prodigiously benefit its college recipients. Through their research, the authors (Taylor is a legal journalist and fellow at the Brookings Institution; Sander an UCLA law professor and economist) found that the current system is actually counterproductive.

“There is now increasing evidence that students who receive large preferences of any kind—whether based on race, athletic ability, alumni connections or other considerations—experience some clear negative effects,” they wrote this week in The Wall Street Journal.  “Students end up with poor grades (usually in the bottom fifth of their class), lower graduation rates, extremely high attrition rates from science and engineering majors, substantial self-segregation on campus, lower self-esteem and far greater difficulty passing licensing tests (such as bar exams for lawyers).”

Sander and Taylor found that leapfrogging students creates an academic and social mismatch that feeds into their feelings of isolation and anxiety. For that reason, they believe it would be much more effective to place disadvantaged teens in a school where they can actually excel.

“The most encouraging part of this research is the parallel finding that these same students have dramatically better outcomes if they go to schools where their level of academic preparation is much closer to that of the median student.”

Specifically, the authors make three main recommendations to reform existing diversity programs:

  1. Applicants who are offered admission should be given information about how other students at their academic level typically fare compared to their peers so they may evaluate the danger of mismatch.
  2. Make sure racial preferences do not exceed socioeconomic factors to help less privileged students of all ethnicities.
  3. Eliminate race-based scholarships and replace them with need-based scholarships. According to a 2011 University of Denver study, American law schools give four times as much grant aid to rich African Americans as to poor whites.

Which brings us back to Allessandra Bradley-Burns. Despite her privileged background, she said in her “Huffington Post” column that the racial preferences bestowed on her were justified because her parents “did not understand or even hold the beliefs of the upper middle class.”

Huh? So now affirmative action should be extended to those whose parents don’t feel rich? That’s nuts, inherently unjust and it clearly doesn’t work as it was originally intended. Taylor and Sander are correct. If the mandate to level the playing field established more than a half century ago is to remain relevant, it has to change. It’s simply the right thing to do.