True, but that was still a biased sample, in that the non-responding 25% probably differs from the other 75% (motivation, wealth/monetary sensitivity). Estimati

True, but that was still a biased sample, in that the non-responding 25% probably differs from the other 75% (motivation, wealth/monetary sensitivity). Estimating a mean depends only on sample size, not population size, which is how a random sampling of 600 people always trumps a biased sample of 6000. E.g., "Dewey defeats Truman".