If, as suggested, these agendas are scrutinised and validated by an open community then all remains well i.e. if someone deliberately chooses to use a set of data to prove their point and they happen to be correct then all is well.
There's a movement in parts of the science community to provide the data along with the analysis. You can incorporate these together and publish them through tools like IPython notebooks.
Instead of waxing philosophically about how people might twist data to suit their needs, it's probably better to be more proactive and teach people how to analyze data for themselves.