Wansink Study on Children’s School Lunch Choices Retracted

JAMA Pediatrics has retracted a 2012 study authored by Brian Wansink, director of Cornell University’s Food and Brand Lab, because of “inadequate oversight of data collection and pervasive errors in the analyses and reporting.” The study, published in the Archives of Pediatrics and Adolescent Medicine, reported that children ages 8-11 were about 30 percent more likely to choose an apple with a cartoon-character sticker over a unbranded cookie, concluding that “brands and cartoon characters” could increase consumption of healthier foods in school lunchrooms.

A reader reportedly sent Wansink a letter in February 2017 noting several errors and the study data was rechecked. In September, Wansink sent JAMA Pediatrics a notice of retraction and replacement acknowledging that the researchers “inadvertently provided an incorrect description of the study design and sample size, used an inadequate statistical procedure, and presented a mislabeled bar graph.” After that notice was published, Wansink said the Robert Wood Johnson Foundation, which had funded the study, notified him of an additional error: the study was conducted on children ages 3-5, not 8-11.

After the study was retracted on October 20, The New York Times reported that “doubts have been cast about other papers” involving Wansink, including a 2012 study published in Preventive Medicine that claimed children are more likely to eat vegetables if parents give the foods a “cool” name. After more than 150 errors were discovered in four of Wansink’s publications, Cornell University implemented new research review procedures but concluded the errors “did not constitute scientific misconduct.”