Also, the pollsters know whether the missing 400 respondents introduced any sort of skew in the demographics or attitudes on politics or health reform (based on responses to the first survey), and they can attempt to correct any such bias with weighting.