Nutrition scientists should exercise constraint in seeking the public eye, says Sander Kersten, newly appointed professor.
The 'thirst for publicity' can, cause researchers, for example, to get into the danger of neglecting to substantiate their arguments. Kersten brings up the ridiculous case of a recent report on chocolate eaters, which described how regular chocolate-eaters gain more energy on the average but are still trimmer. The conclusion reached was that eating chocolate seems to help in losing weight. 'But this contradicts so strongly with our biological knowledge,' says Kersten, 'that it makes us question how such a result could be reached.'
Kersten has been appointed professor of Molecular Nutrition. His research focuses on fats: 'We examine the role of fats at the cell and molecular levels,' says Kersten. 'For example, how are fats broken down, processed and stored in the body.' He also wants to examine the link between the immune and the digestive systems. It appears that these are very closely related.
How do you prevent your research from being tomorrow's coffee-table talk? Kersten says there are three questions to consider.
Is your reasoning plausible?
If you find a connection, check if it has a biological explanation. There must be some kind of underlying hypothesis; don't view the body as a black box. Kersten: 'A result does not appear out of the blue.'
How extensive is the outcome?
Scientists always look for significant results, but neglect the extent of the outcome. Is that ten times as much or just five percent more? Scientists nowadays face the difficulty of coming up with relatively small outcomes nine out of ten times. Kersten: 'Big outcomes would already have been discovered in research in the past decades.'
Does it contradict earlier findings?
Your research is completed and you have a good result, but earlier papers tell a completely different story. It may be better in this case to wait for a big meta-analysis or review to show up which can provide a more definite answer. With the significance standards used currently, there is a one-in-twenty chance that your result is a mere coincidence.