Psychologist: Racial bias seen in state hiring

http://www.desmoinesregister.com

Witness says many tied to state hiring have implicit bias against blacks

About three-quarters of the people able to affect hiring decisions in Iowa’s state government have an implicit bias that makes them favor whites over blacks, a University of Washington psychologist testified last week — a bias that could make them discriminate against African-American job applicants without realizing it.

Professor Anthony Greenwald stopped short, however, when asked in a Polk County courtroom if he could directly prove that bias is responsible for the statistically low number of blacks who win jobs when they apply for positions with the state.

“I would be reluctant to make that contention in the journal publication context, because I think that requires a higher standard of proof,” Greenwald testified. “I can say that it is plausible that implicit bias is a cause of discrimination in the state of Iowa. I regard that as a plausible hypothesis that I would love to test.”

Testimony is about halfway done in the case of Pippen vs. State of Iowa, a four-year-old class-action lawsuit alleging that Iowa’s lack of oversight and accountability in its hiring bureaucracy has allowed discrimination to take root.

The scheduled month-long trial, which enters its third week on Monday, ultimately could be worth as much as $71 million for up to 6,000 blacks who allegedly were denied jobs or promotions since July 1, 2003.

Blacks who seek work in Iowa’s government are statistically less likely to get interviews, less likely to get hired and, if they do get hired, will likely earn less money than their white peers, according to a Rutgers University labor economist who also testified last week. The chance of gaps that size occurring in a racially neutral process, according to reports by economist Mark Killingsworth, is less than two in a billion.

It all sounds familiar to the dozen or so, mostly African-American, spectators who have been monitoring the complicated, statistic-laden trial since it began.

“It reinforces what we already know,” Tereasa Jefferson, a secretary at Iowa Workforce Development, said last week at the courthouse. Jefferson, a pre-2000 state employee who was repeatedly denied job interviews until she landed her current position in 2007, said her main concern is that “I want better for those coming up behind us.”

“We don’t want the moon,” Jefferson said. “We just want to be somewhere in the light.”

The more than 50-volume Pippen court file includes numerous allegations of blacks being denied jobs they were qualified for, of hiring rules being bent in favor of white applicants, and retaliation against blacks who sued the state for discrimination. But the case now being played out before Judge Robert Blink so far has focused mostly on numbers, bureaucracy and large-scale rejection.

Witnesses thus far have included people such as:

Nansi Woods, a Des Moines native and former state employee who wasn’t hired for any of the roughly 75 social work jobs she applied for between March 2007 and December 2009.

Michael Greene, a former U.S. Navy cook and a journeyman plumber who allegedly was denied one state job because a white applicant was allowed to apply after a deadline lapsed. Green, who ultimately was given another interview and worked four months for the state in 2008, now works as facility operations manager overseeing a USDA building in Miami, Fla.

Phillip Wells, who sought a “human services” job with the state but ultimately submitted hundreds of applications for jobs ranging from mailroom clerk to driver’s license examiner in a failed bid to get his foot in the door. “You’re not going to win the lottery if you don’t play,” Wells testified.

The bulk of court time this past week focused on Greenwald and Killingsworth, a Rutgers statistical expert who analyzed five years of state job application data for the plaintiffs going back to 2004.

Killingsworth’s reports, discussed in exhausting, highly technical detail over three days, found that blacks in some years had less than half the chance of getting hired as white applicants.

Iowa attorneys, who have not begun to argue their case in court, previously have contended in paperwork that the plaintiffs are required under federal employment law to prove a link between alleged lapses in state hiring procedures and a race-related cause. Without such a link, lawyers argue, the state could be unfairly blamed for statistical differences that have other explanations.

“As you sit here today, you’re not offering any opinion as to whether any of these differences are caused by any particular practice, are you?” Deputy Iowa Attorney General Jeffrey Thompson asked Killingsworth on Friday.

“I doubt that a statistical analysis is capable of doing that,” Killingsworth replied. “So no, I am not making that kind of statement.”

Thompson’s questions for Greenwald, the University of Washington psychologist, focused largely on the professor’s Implicit Association Test, a test purported to gauge the bias of its taker by measuring the amount of time it takes to sort computerized photos of faces and words with good or bad connotations. (The test can be taken online at the Project Implicit website, https://implicit.harvard.edu/implicit.)

“Just because there’s going to be a bias doesn’t mean people are going to act on that bias,” Thompson said. “For us to walk down your theory, one of the requisites is that there has to be a cue. Otherwise, the IAT data is meaningless, isn’t it?”

Greenwald acknowledged as much.

But in previous testimony, the professor had discussed ways that blacks could be differentiated based on hiring documents and resume lists that include items such as membership in a Baptist church or the NAACP.

Blacks also, on average, are poorer spellers than whites, Greenwald said. And after being shown the resume of Nansi Woods, the professor noted that “African-American names frequently have spellings that are deviant or are totally unrecognizable.”

He later amended that by adding that “ ‘unusual’ would be a better word than ‘deviant.’ ”

Greenwald said biased individuals tend to practice “differential sorting” when faced with such differences. Black job applications may be downgraded for spelling or grammar errors, for example, “where you might forgive it in the case of a white person.”

The trial resumes Monday.