The second question does not ask for an estimate of prevalence, but to pick the answer you think most popular. You can over-estimate the amounts as much as you like as long as the majority choice remains the same, ie it would not matter whether you think there are 0.1% or 49.9% of creationists...