Well this article is couched in terms of the pandemic. It would be interesting to see some studies to tease out political biase media outlits during election season. Does the size of a crowd influence the perception of popularity?
And I'm not talking about sligh misjugdments, I'm talking being off by one order of magnitude. In your typical picture of a protest taken with a smartphone from a high-rise building there is at most 5 to 10k people in the shot. Yet you'll often find such pictures on social medias labeled "100s of thousands in the street protesting for [whatever]".
And for someone who's not particularily paying attention it's easy to be manipulated because we really have no frame of reference for what thousands of people look like, let alone tens or hundreds of thousands.