here's a recent scientific journal article that may be of interest:

Journal: Food Quality and Preference
Volume 15, Issue 2 , March 2004, Pages 91-95

Organic food claims cannot be substantiated
through testing of samples intercepted in the
marketplace: a horticulturalist's opinion

F. Roger Harker,

Abstract

Some studies comparing organic and non-organic foods continue to source products from retailers
(see review by Bourn & Prescott [Critical Reviews in Food Science and Nutrition, 42 (2002)
1]). While the reasons for taking this approach are rarely stated, it is likely that in part the decision
reflects the difficulty in obtaining test samples from agricultural field trials, and the assumption that
mixing of products and/or raw materials during processing and marketing results in an unbiased and
randomized distribution of samples on the retail shelf. However, decisions on appropriate sites,
cultivars, and harvest criteria can differ between organic and non-organic sectors of agriculture.
These decisions do not govern the organic status of a commodity, but may introduce systematic bias
in the quality of food intercepted in the marketplace. Furthermore, the normal distribution of `quality'
obtained at harvest is sometimes modified through the imposition of `quality standards', which aim to
provide the consumer with a higher quality product than they might otherwise receive. Thus `quality'
on the retail shelf can reflect industry regulations as much as the different production systems. This
article uses experiences in the apple industry to highlight how differences between organic and
non-organic fruit observed in the marketplace may be confounded by factors not prescribed within
organic production protocols. Claims that an organic product tastes `different', is `preferred' and/or
`more healthy' have the implicit expectation that the improvement is due to the way the food has
been grown. Therefore, robust experimental approaches should source product from field trials
rather than from retailers.


(Apologies for the line over-runs -- above is how it appears on the
journal website.)

Tom Billings