I've had many people reference the China study and specific individual studies in this discussion that "prove" dairy is harmful.

The first step would be to reiterate that science never "proves" anything, it only disproves hypothesis and through averages makes assumptions on best practices within a highly specific context.

That being said, there is a hierarchy in research ranging from observation and opinion to meta-analysis and systematic review. Near the bottom of the hierarchy are epidemiological studies or uncontrolled longitudinal studies. The latter are what make up the China study. Almost entirely correlative data that have been invalidated at higher levels of scrutiny.

The distinction between correlation and causation for instance is neglected throughout the china study. The article Why correlation does not imply causation? does a good job at breaking this down. And the research (although abundant) is cherry picked to support a preexisting agenda or philosophy. These studies are not representative of the large body of research literature on the subject that suggest the high biological value of dairy products.

To further elaborate on my point that correlation does not equate to causation lets look at another correlation in exercise physiology. Cortisol is the hormone that is most strongly correlated to hypertrophy. However, we know that cortisol is catabolic in nature with regard to muscle tissue and has no net positive effect on anabolism post workout.

The correlation is in the consistent presence of cortisol post workout, but the hormonal signalling responsible for muscle anabolism is entirely different. Strenuous exercise is catabolic in nature and causes a stress response which stimulates production of cortisol. This however is followed by a hormonal cascade that induces hypertrophy through various metabolic machinery that is entirely unrelated to cortisol.