The assumption that people lie on surveys is the same as the assumption that they don't. Equally valid. People will tell a survey whatever suits them most.
People tend to be truthful more than untruthful. This is kind of hardwired into humans, due to mutual trust requirements for a successful society.
This is not the same as saying that nobody ever lies on a survey only that an assumption of truthfulness is more likely to be correct than an assumption of untruthfulness.
In any case the problem is that John/Argus assumes truthfulness where it suits him and untruthfulness where it suits him. This says more about him than aboit the survey respondents.