This kind of bias is a statistical measure; typically you can't prove or disprove it using a single sample.

It's about larger patterns, which things are talked about and (crucially) which are not. How much attention is given to things and not.