As usual, it's important to point out that this was not an experimental study design (there was no random assignment to treatment and control groups). Peter Norvig's online article about what to look for in research studies
I was going to mumble something about correlation != causation, but I decided to track down the paper first (http://bjp.rcpsych.org/cgi/content/full/195/5/408). The authors didn't explicitly decide that diet caused depression, but they did note that the data seems to invalidate the idea that depression caused bad diets.
I've done a lot of writing and analysis in health sciences. My very strong intuition is that there is a study flaw here of one form or another.
What this seems to say, to me, is that people who choose to eat healthy also tend to be less depressed than those who choose to not eat healthy.
Population bias may be a problem here - that eating unhealthy may not necessarily cause depression, but rather those who are eating unhealthy are also likely to be suffering from depression.
There is limited data on diet and its impact on health in general - except in obvious cases of frank deficiency.
This could be like saying "People who go jogging daily tend to be happier in general." You could say that jogging then causes happiness. Or you could say that people who tend to jog are demographically likely to also do happy inducing things.
They're accounting for gender, age, education, and physical activity, but are they also taking into account things like free time and stress level and salary? If I need shove junk food into my face between my day job and my night job, of course I'll be more depressed than if my butler (or non-working spouse) makes me a parsnip soup after my 10-4 shift.
The shareholders of food processing companies like it. The reason we have so many processed foods is that traditional "whole food" businesses are not good businesses to invest in. The population is not growing and people do not need to eat more calories, so that means that your revenue is not going to grow year over year, because the US basically needs the same amount of food every year. This means that you have to make money by using less raw material and charging more for it; hence, processed foods.
(Sure, Store A may steal sales form Store B in the short term, but in the long term, the industry will not grow.)
I've started tuning out every study that purports to "link" two things, particularly if the claimed effect is cancer or the claimed cause is chic to hate on.
Yes. There should have been professional follow-up on this by now, and it's odd that a BBC story from more than a year ago is the submitted link for this study.
[+] [-] tokenadult|15 years ago|reply
http://norvig.com/experiment-design.html
suggests other issues to examine here.
[+] [-] mkross|15 years ago|reply
[+] [-] gohat|15 years ago|reply
What this seems to say, to me, is that people who choose to eat healthy also tend to be less depressed than those who choose to not eat healthy.
Population bias may be a problem here - that eating unhealthy may not necessarily cause depression, but rather those who are eating unhealthy are also likely to be suffering from depression.
There is limited data on diet and its impact on health in general - except in obvious cases of frank deficiency.
This could be like saying "People who go jogging daily tend to be happier in general." You could say that jogging then causes happiness. Or you could say that people who tend to jog are demographically likely to also do happy inducing things.
[+] [-] lsb|15 years ago|reply
[+] [-] vidar|15 years ago|reply
[+] [-] jrockway|15 years ago|reply
(Sure, Store A may steal sales form Store B in the short term, but in the long term, the industry will not grow.)
[+] [-] aces|15 years ago|reply
http://www.demotivate.info/wp-content/uploads/2008/08/behang...
[+] [-] haberman|15 years ago|reply
I've started tuning out every study that purports to "link" two things, particularly if the claimed effect is cancer or the claimed cause is chic to hate on.
[+] [-] PonyGumbo|15 years ago|reply
[+] [-] silentbicycle|15 years ago|reply
2009 is hardly ancient history.
[+] [-] tokenadult|15 years ago|reply
Yes. There should have been professional follow-up on this by now, and it's odd that a BBC story from more than a year ago is the submitted link for this study.