Wait a minute, didn’t we already have the grapefruit diet, the avocado diet, etc? And who could forget the Atkins Diet with its rashers of bacon? Does everything qualify as healthy food?
Researchers offered one group of kids raisins and grapes as after-school snacks; the other, chocolate chip cookies. By the end of the day, the raisin diet kids wound up eating fewer total calories.
Science Speculation on the Raisin Diet
In this case, rather than “speculating” – going beyond the news – it seems more appropriate to dig deeper, to get behind the headlines. Facts are great, but they don’t tell the whole story, and this news story is a great one to demonstrate skeptical reading. A few things I like to look at when reading science-related news:
– Who paid for the work?
In this case, the California Raisin Marketing Board sponsored the Journal of Food Science issue that published this work on the raisin diet. I always assume that researchers are honestly reporting their work. But even with honest researchers, there are many ways that money distorts results: The sponsor got to choose whether or not to pay for the project, knowing what the researchers proposed to do. The sponsor may have had the right to delay or block publication if the sponsor didn’t like the results. And projects that might undercut the sponsor’s marketing program probably didn’t get funded at all.
– Causation versus association.
Causation versus association is mainly a problem with demographic studies. (That is, Japanese men eat lots of rice. Japanese men have a lower incidence of prostate cancer. Therefore, eating lots of rice is associated with less prostate cancer, but may not cause less cancer.)
In the case of the raisin diet study, there was a control group, so eating the raisin diet may in fact have caused the kids to eat fewer total calories. But it’s also possible that those kids who chose to eat more raisins and grapes came from homes where parents discouraged eating to excess. In other words, the results of the raisin diet study could have been caused by outside factors not under control of the researchers.
– Is there a plausible model?
If the researchers don’t offer a reasonable explanation for the result, then we should always worry that something has been forgotten, something that might change the conclusions we draw from the study. Can we invent an explanation? Well, maybe. Raisins have a glycemic index of 64, grapes 46, chocolate chip cookies 64. So perhaps the grapes don’t cause your blood sugar level to bounce and crash, making you hungrier. So raisins may indeed be healthy food. But it’s not obvious that they reduce your weight.
– Is the study well designed?
This is a tough one. News articles rarely discuss the experimental design. It would be cheating to simply gather a great deal of data and then search through the data — “data mining” — to find associations and relationships. However, judging whether the research project was carefully designed to avoid more subtle kinds of data selection requires studying the original technical publication, and often requires expertise in the field. When the original article leaves me confused, I look for a comprehensive review article, where a recognized expert surveys many research studies and tries to make sense out of them.
Is your family ready to try the raisin diet? What suggestions do you have for critical reading of stories like this one?
http://www.upi.com/Health_News/2013/06/28/Raisins-may-lower-food-intake-help-prevent-diabetes/UPI-68641372454525/, Raisins may lower food intake, help prevent diabetes;
http://onlinelibrary.wiley.com/doi/10.1111/1750-3841.12070/abstract, An After-School Snack of Raisins Lowers Cumulative Food Intake in Young Children
https://www.eurekalert.org/pub_releases/2013-08/ngaw-gca080213.php, Sacramento Study on Grapes and Raisins as a Healthy Food;
http://onlinelibrary.wiley.com/doi/10.1111/1750-3841.12066/abstract, Improved Diet Quality and Increased Nutrient Intakes Associated with Grape Product Consumption by U.S. Children and Adults: National Health and Nutrition Examination Survey 2003 to 2008