The devil is in the details… ‘who’ was surveyed? I’m sure I could survey a ‘picked’ group of people and get a positive reading of ‘whatever’ result I wanted.