Defense Cites Faulty DNA Calculations : Simpson trial: Expert witness for the prosecution admits he overestimated the probability that bloodstains on glove and Bronco are the defendant’s.
Struggling to regain his credibility after admitting faulty calculations, statistician Bruce Weir acknowledged Monday that bloodstains on O.J. Simpson’s Ford Bronco and a crime scene glove contained genetic markers far more common than he had previously estimated.
A smear on the Bronco’s console, for example, contains markers consistent with Simpson--and also one in 570 other people. Last week, Weir had testified that only one in 1,400 people could have been a source of that blood, which contains several DNA types.
In a tough cross-examination, defense attorney Peter Neufeld tore into Weir, capturing jurors’ interest with repeated suggestions that the prosecution’s star DNA witness was unreliable.
The prosecution has relied heavily on the DNA evidence to connect Simpson to last year’s slayings of his ex-wife, Nicole Brown Simpson, and her friend, Ronald Lyle Goldman.
Simpson has pleaded not guilty. And on Monday his attorneys tussled with prosecutors about the validity of DNA evidence. The two sides offered strikingly different statistics about how many people in the world might share genetic markers found in bloodstains in Simpson’s Ford Bronco and on a battered leather glove.
At the end of the day, prosecutors shifted from the dry world of statistics to the horror of the crimes. Presenting evidence about hairs and fibers found at the crime scene, they flashed gory photos of Goldman’s clothing: a white shirt as drenched with blood as a butcher’s apron, a pair of jeans so deeply stained they appeared to be two-toned.
Seemingly relieved to be presented with concrete evidence once again, jurors paid careful attention to the photographs. Even bland shots of the envelopes containing Nicole Simpson’s fingernail clippings and Goldman’s light-colored, medium-length hair interested them.
But much of the day was devoted to far more abstract topics. Even statistics professor Weir testified with a wry smile that most people have trouble grasping the big numbers associated with DNA analysis.
Yet both defense attorney Neufeld and prosecutor George Clarke plodded on with questioning. Neufeld’s sharp, insistent grilling seemed to resonate more with jurors.
Seeking to minimize the DNA evidence, Neufeld told the jury that 10% of the population could be responsible for a smear mingling three different blood types on the infamous glove. But Weir insisted that the defense was using incorrect, misleading formulas.
His calculations, he said, showed that only one in 1,600 people--including Simpson--has genetic markers consistent with blood on the glove.
Weir came up with the statistic in a private number-crunching session over the weekend. He was forced to scrap his original calculation that just one in 3,900 people might match the blood after the defense team last week pointed out that he had botched the math.
Clearly embarrassed, Weir returned to the stand Monday and announced that the mistake was “worse than I thought"--the same flawed assumption ruined several of his calculations. Under a tough defense grilling, he acknowledged that his error had made the blood markers look at least twice as rare as they really were.
“I’m going to have to live with that mistake for a long time,” he said.
Later, he conceded that he had not subjected his new calculations to a rigorous cross-check that he had boasted about just last week as a near-guarantee that his statistics would be accurate. “I had to sleep sometime between midnight and 5 this morning,” he said.
Certainly, Weir got little rest on the stand.
Neufeld probed for any weaknesses in the statistical evidence. He questioned how Weir could estimate the frequency of genetic markers in the population at large by studying a database that contained no Asian Americans and no California residents. He called Weir’s mathematical formulas experimental.
And he zeroed in on the familiar defense theme of laboratory error: Fancy calculations are meaningless, he suggested, if they are based on flawed DNA analysis.
Weir fiercely defended his statistics, arguing that he could not and should not consider potential lab errors.
But Neufeld turned the witness’s own words against him, reading out loud Weir’s testimony in a North Carolina case last month in which he agreed that jurors should be told about labs’ error rates when weighing DNA evidence.
The jurors, some of whom had been fighting to stay awake, seemed to shake their boredom and take note.
In a marked change from his condescending tone last week, Weir seemed to be trying to better connect with the jurors, turning to face them several times. Still, he continued to stumble. Twice he misread his own charts and had to correct himself--prompting several jurors to roll their eyes and shake their heads.
And Neufeld seized every opportunity to undercut Weir’s testimony. It was a familiar defense strategy: Over the last few months, Simpson’s lawyers have portrayed Los Angeles Police Department detectives, an LAPD criminalist and a Los Angeles County deputy coroner as bumbling, unreliable and even incompetent.
Again and again, the attacks put prosecutors on the defensive. On Monday too, Clarke had to spend much of his time trying to explain away mistakes and resurrect his own witness.
Explaining that he was not being paid to perform calculations for the prosecution, Weir said he had agreed to testify because “I’ve been very angry about some of the statements about DNA statistics that have been made in court [over the past few years.] I’m very anxious that the [analysis] be done correctly.” A statistics professor at the University of North Carolina, Weir said that while he was testifying in Los Angeles, he was officially “on vacation.”
Preparing New Figures
Weir churned out 300 pages of numbers over the weekend as he revised his erroneous calculations. Court was delayed for more than an hour Monday morning as defense attorneys pored over the new math.
When the jurors filed in at 10:15 a.m., Neufeld fiercely attacked the validity of the population samples used to determine how many people share certain genetic markers.
Weir acknowledged that the most accurate way to figure out the frequency of genetic patterns would be “to go out and type everybody in the world.” But he said such precision was impractical: “We can’t count everyone in the world, much less type them.”
Neufeld suggested that at the very least, databases should include people from the “population of potential perpetrators.” In this case, Weir acknowledged, that population would include residents of Los Angeles County.
Instead, the database included several hundred FBI agents and blood donors from Detroit, Miami and Houston.
“Do you consider 220 white FBI agents to be in the population of potential perpetrators for the murders of Ronald Goldman and Nicole Brown Simpson?” Neufeld asked, with heavy sarcasm in his voice.
Weir responded: “Oh, yes . . . it’s a lovely collection.”
In seeming disbelief, Neufeld noted that the population samples contained no Asian Americans.
But Weir insisted that for geneticists, the concept of race is meaningless. “We may look different,” he said, “but underneath, we’re all the same.”
Jurors, looking bemused, listened with interest as he added: “The whole idea of race is very vague. . . . People can call themselves Martians, and we’ll still get the same results: that the profiles [found in crime scene blood] are rare.”
Nonetheless, Weir did acknowledge that a given genetic marker might show up more frequently among a certain racial group.
The most likely source of a mingled blood spot on the Bronco’s steering wheel, he said, was a trio of two African Americans and one Caucasian. He pegged the chance of that combination leaving the stain at one in two. In contrast, the probability that three Caucasians were responsible for the blood smear was just one in 3,500, Weir said.
Groping for a common-sense metaphor to explain the tangle of statistics, Weir compared the analysis of DNA frequencies to a slot machine.
Imagine, he told jurors, a stain consisting of three types of blood. To hit the jackpot, the “slot machine” of DNA analysis must come up with a lemon on one wheel, a cherry on another and an orange on the third, he said.
Weir said his statistics calculated the frequency of hitting the jackpot--of three people possessing the right genetic markers to match the mingled elements in the drop of blood. Neufeld’s much lower odds, he said, reflect the chance of either a lemon or a cherry or an orange popping up on the slot machine--not of all three flashing simultaneously.
Thus, when Neufeld announced to the jury on Monday that half of humanity could have left the stain on the Bronco’s console, Weir dismissed that statistic as “absurd.” His own calculations showed that only one in 570 people could be responsible for the blood smear. And O.J. Simpson could not be excluded as a possible source.
“I wouldn’t even dignify [Neufeld’s] number by putting it on the same paper” as his own figures, Weir said. “I’m not saying it’s wrong, it just is wrong.”
Trace Evidence Presented
When Weir finally stepped off the stand, the prosecution embarked on the next phase of their case: a discussion of hairs, fibers and other trace evidence.
LAPD crime lab technician Denise Lewis described, in detail, how she changed her latex gloves before handling every piece of evidence, from soil to fingernail clippings to a moldy, bloody shirt.
After establishing the chain of custody for those items, prosecutors plan to call two other witnesses to discuss the evidence. Outside court, Deputy Dist. Atty. Christopher Darden said the prosecution hopes to wrap up its case by Friday.
When court adjourned at 5 p.m., Superior Court Judge Lance A. Ito asked two jurors to hand over books they had been reading so he could check them for “problem” material. One of the books reportedly was a John Grisham novel.