UW battle over Baby Einstein settled, maybe
The University of Washington has handed over $175,000 and the raw data supporting a research study to the creators of the Baby Einstein products, settling a long-running battle over access to the data.
Seattle Times staff reporter
Four years ago, a University of Washington study largely debunked the educational value of early-childhood education videos, showing that products like Baby Einstein might slow, rather than enhance, language development of very young children
The study went viral in the media and in parenting blogs. The UW heavily promoted it, and the researchers went on the "Today" show.
Disney, which by then owned Baby Einstein, eventually offered a full refund to everyone who bought those videos from 2004 to 2009, and it dropped the claim that the videos were "educational." It cost the company an estimated $100 million.
What went largely unnoticed was a long-running legal fight between Baby Einstein's founders and the UW, which was settled this month with the UW handing over $175,000 and the raw data supporting the study to Bill Clark and Julie Aigner-Clark.
Clark said he wanted the data to see if the study could be replicated and to restore his and his wife's legacy. But his quest was bogged down by competing concerns between the state's broad mandate for open public records and the privacy rights afforded to human research subjects. Now that he has it, he says, a series of "irregularities" has emerged and that goal is impossible.
"I can't trust any of the data they've given me," said Clark, who with his wife founded the company in 1996 and sold Baby Einstein to Disney in 2001. "The idea of reproducing their study with certainty is out the window."
During the public-disclosure lawsuit, filed last year in King County Superior Court, the UW revealed that the lead researcher ordered the underlying data — telephone surveys with more than 1,000 parents — destroyed sometime in 2008, apparently violating UW policy. A backup copy was later found, according to documents filed as part of the lawsuit.
The UW had denied several of Clark's public-disclosure requests for the data since 2007. Attorneys said that was in part because it could not figure out how to disseminate the data without violating privacy rights guaranteed to human research subjects.
But faced with litigation, and potential fines for withholding public documents, the UW in 2009 provided to Clark a 2,000-page printed copy redacted to hide participating children's ages and other identifiers. Earlier this month, that material was provided in a fuller electronic version as part of a final settlement.
Those two versions, however, do not match up, Clark said. The discrepancies appear in critical areas of the database, related to how many words children recognized after watching videos such as those produced by Baby Einstein, he said.
In a letter sent last week to Clark's attorneys, UW lawyers explained that the data discrepancies possibly were caused by "software or hardware glitches" in the process of converting the massive electronic database to a printable format.
UW spokesman Bob Roseth said the data discrepancies are "obviously regrettable in a process like this," but that the university stands behind the study. "We think the conclusions in the paper are supported by the underlying data," said Roseth.
Roseth said administrators were "unhappy" that the lead researcher, Dr. Frederick Zimmerman, had ordered the data deleted.
Although the UW requires data to be retained for at least six years, the deletion request is not considered a violation of policy because the information was eventually found.
"It's a violation only if the data is irrecoverable," said Roseth. "Even if (Zimmerman) made an effort to destroy it, we maintained a data set that is exactly the same."
Zimmerman, who left the UW in 2008 and is now chairman of the UCLA Department of Health Services, did not respond to interview requests by email or telephone. His automatic-reply email said he was out of town through July 4.
The study, published in the Journal of Pediatrics, found that for every hour a child 8 to 16 months old watched educational videos, they understood six to eight fewer words than their peers.
Dr. Dimitri Christakis, a doctor at Seattle Children's hospital who co-authored the study, said other researchers have duplicated their findings.
"[Clark] doesn't like the research findings and has been out to discredit them," said Christakis. "I don't have a commercial interest in this. If there is a non-disinterested party here, it's clearly him."
Christakis said he was "very frustrated" by the discrepancy between the printed and electronic versions of sets, but there was no manipulation. "There was a single data set in possession of a person at the university," he said.
The lawsuit, pressed by Clark's attorneys, including Greg Overstreet of Olympia, became bogged down by the patient-privacy concerns.
Karen Moe, director of the UW's Human Subjects Division in the Office of Research, said she was satisfied with the settlement, hammered out in May and June, which provided the children's ages, but no identifiers such as name or location. "I certainly feel the identity of the children and their families have been protected," she said.
The $175,000 settlement pays for most of Clark's legal fees. He said he pursued the issue in part to clear the legacy of his wife, a former teacher who has Stage 4 breast cancer. He said he will try to reconcile differences in the data sets, and is pressing the UW to launch an internal investigation.
"They need to reproduce the data discrepancies, or have a plausible excuse, or they need to stare at the possibility of misconduct," said Clark.
Jonathan Martin: 206-464-2605 or firstname.lastname@example.org
Career Center Blog
Dig into local Gardening