How to test hypotheses about social media manipulation

How could researchers go about testing the hypothesis that social media is showing content that puts us into a state of threat and anxiety? I offer some thoughts and also offer some critiques of these methods.