The problem is that while there are lots of observational or cross sectional studies on soda and weight gain, there are very few studies which, by their design, can show cause and effect.
FORBES: ABC News Attacks Scientist Who Exposed Bias In Obesity Research