The most important thing she points out is this: "The analysis also found that the majority of studies examined did not adjust for factors such as smoking, which is a major risk factor for cancer. If, for example, people who smoke are more likely to dye their hair (or vice versa), than the results would have to be interpreted differently."
This is basically the same problem the "Bacon causes cancer" study suffered from - not adjusting for factors like exercise, smoking, other foods you end up with a correlation study when what people want to see are causation studies. Unfortunately there's such a dearth of critical thinking/scientific skills and such a "need" for sensationalized headlines that we end up with these "common wisdoms" taken from half-studies...
Look for yourself, and you'll see that just about every page that warns of dangers of hair dyes and cancer (bladder cancer in particular) references either the one discredited UCLA study from a while back, or one single "homeopathic doctor".
