Arguably, all of them are at least somewhat biased, but the most biased would be <span>surveying people 15 to 20 years old leaving a concert </span> because being at a concert would be highly correlated with liking the type of music played at that concert.
