"Britons are eating 50 per cent more food than they admit, according to the first official use of a biological test to measure calorie intake. The scale of people’s lying and inaccuracy about what they eat undermines dietary recommendations based on shaky research, experts said. The findings show that the average person in Britain is exceeding official health recommendations by the equivalent of a Big Mac a day. Men are consuming 3,119 calories a day, not the 2,065 they own up to; women are consuming 2,393 instead of the 1,570 they confess to, the Office for National Statistics (ONS) estimates." https://www.thetimes.co.uk/edition/news/the-big-fat-lie-britons-eat-50-more-than-theysay-cpxcgnh8p I thought this was interesting in the context of self reported questionnaires. As a rule of thumb you kind of assume a bias of 10-30%, but this shows that when it comes to some things the bias can be a lot higher. I would suggest that social stigma is driving this. If this is true, then bias on questionnaires would be expected to be higher with subjects that are tainted by the more 'socially unacceptable' stigma such as perceptions around greed (calorie consumption, eating healthily, alcohol consumption) or laziness (amount of exercise etc) than something more benign such as getting enough sleep or water consumption.