I'll preface this by saying I didn't read the article, nor did I read any of the studies and underlying methodology so it has probably been addressed and corrected for but like a few of the other commentors have mentioned, by measuring it based upon consumption of a single item, it would be hard to see if it really just showed an indicator of overall consumption as opposed to a singular food being the cause.

Lets say one of our sample respondents consume 350g of red meat on average in a week and that consisted of approximately 10% of their diet (by weight). Compare that to a person who had 350g of red meat on average in a week and it consisted of approximately 5% of their diet (by weight). This would be an Extreme example but the second person is literally consuming twice the amount of food (by mass).