Comment by kar5pt

Their methodology seems really flawed. They cherry picked a thousand instances of wrongly removed content, but how many posts do you think there are about the conflict on Facebook and Instagram? Tens of millions? They have no way of knowing how representative that sample is.