<p>My experience at Williams suggests that there is considerable polling and survey expertise on campus.</p>
<p>I don’t know which survey instrument Williams used last spring (the CSEQ survey, the ESS survey, or one of the COFHE surveys), but all of these are nationally available commercial surveys, specifically engineered for higher education customers, that have been administered at places like Williams and virtually every other school on a regular basis for decades.</p>
<p>This is not a tricky survey challenge because you aren’t dealing with a small representative sample. You are surveying the entire population with a very high response rate - typically 50% to 60%. Don’t blame the survey.</p>
<p>The reason the percentage of Asian-Americans and African-Americans matters is that, nationally, both student groups binge drink at half the rate of white students. Thus, for every white student you replace with an Asian-American or African-American student (as Williams has done over the last five years), the binge drinking rate is statistically likely to decline.</p>
<p>Similarly, athletes tend to binge drink at higher rates than non-athletes nationally – findings that have been consistently supported by both recent Williams surveys.</p>