The story of Chocolate and Intentions was highlighted in this months Mini-AIR [from Improbable Research] which I am including here:[Image]
2009-02-09 RESEARCH SPOTLIGHT: Thought-Laced Chocolate
This month's specially selected study describes an attempt to blend intention into chocolate, and to then measure the effect upon individuals who consume the hybrid. The paper is: "Effects of Intentionally Enhanced Chocolate on Mood," Dean Radin, Gail Hayssen and James Walsh, Explore, vol. 3, no. 5, September 2007, pp. 485-492. (Thanks to investigator Mary Beckman for bringing this to our attention.)
The authors, at the Institute of Noetic Sciences in Petaluma, California and at Hawaiian Vintage Chocolate in Honolulu, Hawaii, explain: "A double-blind, randomized, placebo-controlled experiment investigated whether chocolate exposed to "good intentions" would enhance mood more than unexposed chocolate....
Each person consumed a half ounce of dark chocolate twice a day at prescribed times. Three groups blindly received chocolate that had been intentionally treated by three different techniques. The intention in each case was that people who ate the chocolate would experience an enhanced sense of energy, vigor, and well-being. The fourth group blindly received untreated chocolate as a placebo control.... "
Conclusion: The mood-elevating properties of chocolate can be enhanced with intention."
A copy is online at http://www.deanradin.com/papers/chocolate.pdf
The story continues at Improbable Research, where Poul Robinson offered this follow up:Picking up where this left off, I thought I might add to the critique:
One of the main conclusions, analysis #3, presents a one-sided p-value of 0.04. One-sided p-values are definitely not the standard in medical research, and this p-value should be doubled (~0.08 two-sided), and this interaction considered to be non-significant.
Analysis #5 does not seem to depend of a one-tailed p-value. However, I note that the confidence interval widths for the control group at day 5 is roughly 10 in both figures 4 and 7, despite there being fewer subjects for the “chocolate sensitive subset”. This is not impossible, as the subjects who contributed most to the variance might also have been those who consume the most chocolate, but it is curious.
Also, the pattern of mood disturbance scores for the control group between days 3 and 7 exhibits a distingctive “peaked” shape, somewhat reminiscent of a Hershey’s Kiss. [Image]
Legal Disclaimer: The staff at DTA would like to remind you that chocolate, like tomatoes, can be very hazardous, and should only be consumed by properly equipped and supervised professionals. We cannot be help responsible for the possible effects of chocolate, which include but are not limited to, romantic relationships, happiness, and eventual death.[Image]
Happy Valentine’s Day from Dread Tomato Addiction (just a little early).
One advantage of Bayesian analysis is that you can but a small prior on nutty results and get a small posterior unless the data are strong. My prior probability on the null for this chocolate study would be very large.
ReplyDeleteSee Musicians, Drunks, and Oliver Cromwell for examples.
Very large indeed, but somehow I think we might still end up with nuts in our chocolate. ;-)
ReplyDeleteA non-parametric analysis might also be interesting, given that the variance is "curious".
So Bayesian analysis can get one a small posterior through the consumption of chocolate? Now that deserves further study!
ReplyDelete