I'm not sure I understand what the motivation is behind lying to one volunteer survey, then not lying to another volunteer survey?
It's all about the way the question is asked, as usual. If people know that the survey is about religious attendance, they will give the answer they think is right or expected of them, rather than the correct answer. If you ask people a more general question about what they did step-by-step last weekend, they'll just tell you the truth. Well, if they went to a strip club or something, they might leave it out, but basically, if you explicitly mention religion in questions, you tend to over-report the religiosity of a group. It's no secret that the number of Christians in the UK plummets from over 70% to under 50% when the question asked isn't the sort of loaded question we see in the census, for example.