but why don't they make sense? you haven't alluded to it at all.

as you said before, that is just an example. i'm referring to the use of decoy questions period. you said decoy questions are irrelevant; i said why they are not.

sorry i don't understand what you mean when you talk about the numbers and publishing. 400 is also a small number but apparently that's enough for you to conclude that "a large majority of women" try to become pregnant without their partners knowledge.

what i meant was that if you sent out 5000 surveys, and 500 people responded, you would say we surveyed 500 people; i think the average reader then reads that as 500 people responded in this way and thus is an accurate representation of a population. reporting the response rate (surveyed 5000 only 500 responded) then demonstrates the gap created by non respondents which indicates how reputable the information is. to go back to your example, one-third of 400 women surveyed said that they risked pregnancy. however, if more people responded (and since there is no report rate, we don't know how many people didn't respond) we would have totally different statistics. one-third of 400 is approx. 130, while 1/3 of 5000 is approx. 1666.

another thing we've completely neglected is the demographics of the group interviewed. 400 community college students. i think your results would vary wildly if you took a more generalized survey.

again i don't understand what you mean by biased against themselves. a good survey would present both questions (have you cheated and have you been cheated on). this comes back to the decoy questions. if i want to survey cheaters, i will also ask them whether they have been cheated on so they are more likely to answer honestly.

what exactly is nonsense? i think i have been very clear and tried to explain where you pointed out that you were confused.