A disproportionately large number of
Why this is right
When we see surveys, we want to stay aware of the possibility of response bias. One of the options we came up with with is that people who responded really were in the top quarter and were just more likely to respond than other people in the class. They would be disproportionately represented in the survey. Let's use actual numbers to draw this out with a class of 200. Class of '60 (25% top ranked): Total: 200 students Top Quarter: 50 students Survey (50% top ranked): Total: 100 students Top Quarter: 50 students In the whole class, there were 150 lower ranked students. In the survey, there were only 50 who responded. If all of the top ranked students responded to the survey, they could have been telling the truth but would take up a bigger proportion in the survey than in the original class.
Skill tested: Paradox · how this choice captures the argument's function is the move to repeat next time.