This was the first thing that I thought of also - surveying Stack Overflow survey-takers about their place in the larger order of programming threatens sample bias in several ways (experience, community engagement, etc).