Census religion question must be made clearer
I have a related concern that the census questions should be drafted properly in order to increase the accuracy of the answers given. As Mr FitzHarris suggests, the last census clearly overestimated the number of Roman Catholics in the country, and underestimated the number of people of no religion.
This was not, however, due to dishonesty on the part of the surveyed. The question on religion in the census seemed designed to have this effect.
