data dredging

Is being overweight in childhood linked to mom's prenatal caffeine intake? If it is, it's not terribly well supported by this new study published in BMJ.
We ve written before on scientific fraud and the problem of how easy it is to get papers with fake or manipulated data published. These studies that somehow make it through the publishing process can range from relatively harmless, such as the deliberately faked chocolate is good for weight loss study,
Why are there so many scary stories about common foods and products causing cancer? An excellent blog-essay explains quite a lot about how science (or science-y) journalism works. It may confuse or scare you, but not as much as the so-called research does.