I think most papers only seem contradictory because: 1) the conclusion has been simplified to the point where the meaning is lost (e.g. "coffee is good for you

I think most papers only seem contradictory because: 1) the conclusion has been simplified to the point where the meaning is lost (e.g. "coffee is good for you!" vs "black coffee is marginally better than soda, red bull, highly processed fruit juice and all the other junk people typically drink." This is very, very common) 2) the study has significant limitations (e.g. we just looked at what people ate and how healthy they were, and tried our best to identify and account for all the many confounding factors after the fact. Epidemiological studies can be useful to identify potential areas of further research but are usually not very strong evidence by themselves) 3) the study was designed to come to a conclusion that a vested interest wanted (e.g. most people have different natural cholesterol "set points" given the same diet. "So if we look at the correlation between egg consumption and cholesterol in a small group, instead of the change in cholesterol for the same people eating eggs vs not eating eggs, the effect will be lost in the noise and we can make it seem like eggs don't raise your cholesterol significantly", gleefully thought the American Egg Board) I'm a big fan of NutritionFacts.org for breakdowns of all sorts of studies, pointing out all the limitations and figuring out what the studies actually say.