I don’t trust any article or survey that compares how many people ate at a restaurant when restaurant dining was closed and uses this number to try to convince people that behaviors are changing. Since this was mentioned in the article and not confirmed to be an invalid measure, I trust none of it.