But would you also engage in these behaviors if you couldn't go online and A) project action onto an "army" that neither exists or has occurred; and B) gain social prestige by allying with a dominant narrative?
The question isn't intended to be rude, or suggesting your motives are tarnished. Just a query into the extent online projection (and presumptions about it) affect our choices. Note also that I am distinguishing between your stated choices and the choices themselves.
Before that data point resolves itself, remember that n=1 doesn't really support either meta-narrative. Not saying you are doing this, but I'm sure there are many who are preparing to both sharpen this data into a spear for the "other side", and at the same time, preparing a cognitive dissonance process to explain the data should it not support their meta-narrative.
The end result is discussion that advance little in the Shannon information sense, but maximally clarify every individuals social net.
Someone out there is doing a study on meta-communication during COVID. The internalization of dominant narratives on social media is such an unspoken axiom, that merely SUGGESTING optimism results in a flurry of "we can't be sure!" finger-waves.
One could hypothesize that those who derive greater self-esteem or self-worth from social dynamics would engage in behaviors where the long-run probability of increasing self-esteem is higher. In other words, the most adaptive posts on both sides of a social media debate would be those which: A) maximally increase self-esteem via ingroup loyalty, and B) cause the debate to be prolonged or not resolved, thereby increasing the duration of A. The "we can't be sure" defense in response to optimism is B.