First of all, and with all due respect to Noel, I wouldn't be posting a survey 'just for fun' in the first place, because the statistical information which can potentially be gleaned from them is so useful. And with that in mind, multiple surveys with biased questions will only end up presenting a fragmented view of what the modelling community wants. In addition, stats from such surveys can easily be distorted because a certain amount of people with no interest in the specific area(s) being dealt with will still take the survey and submit potshot answers.
I'm not trying to be argumentative here, I'm simply wondering why someone would go to the effort of creating a survey and collating the results without realising the full potential of the survey.