NC State’s women’s basketball team is on the rise. At 22-7 on the season, ranked #21 in the country, fans are rightfully looking forward to seeing the Wolfpack back in action during the NCAA tournament. Before March madness begins, however, the ACC tournament awaits. What if State were to play UNC; who would win?
State already dominated the Heels twice this year. First, the Pack beat down UNC in Chapel-Hill back in January, 66-53. In case there were any further questions, the Wolfpack demonstrated who’s boss in Raleigh, by the score of 73-54. Ouch.
PackPoll wanted to know; who did students think would win if the two teams met again in the ACC Tournament? So, we asked them! Apparently, 4% misunderstood the question, because they answered the Tar Heels would emerge victorious. Another 30% were simply too polite to answer the question. Bless their hearts. The rest, the 66%, the best and the brightest, yeah – these students aced this quiz. Pack you very much.
But we were curious. Who were the polite students that refused to acknowledge State’s destiny? Who were these poorly informed students that couldn’t see the third “W” coming? Or maybe it was the work of Russian trolls? Fake News? The Daily Tar Heel; et tu? So, we dug a little deeper.
Nah, mainly it’s just silly graduate students apparently too busy trying to solve problems like beating the crap out of cancer. While 71% of undergraduates sensed an inevitable Wolfpack victory, barely half (52%) of graduate students also answered correctly. To be sure, not many grad students were depraved enough to say UNC would win – just 7% claimed that would happen. Yet, fully 41% of these intergalactic creatures couldn’t predict a winner, ostensibly because they were too busy studying up on the latest fashion trends or responding to “strictly personals” on Craigslist. For real.
Lady Ballers, should you meet the Heels again in Greensboro, Pack Poll wishes you good luck, even though you don’t need it. WOOF!
NOTE ON METHODOLOGY: This PackPoll took place February 22-25, 2018. We administered the survey over the internet to a random sample of 3,500 NCSU undergraduates and 3,500 NCSU graduate students. The true population of undergraduates and graduate students taking at least one face-to-face class on campus is 21,192 and 7,113, respectively. However, 44 students are minors and 814 more have privacy flags preventing contact, slightly reducing the population eligible to be sampled. Each student remaining in the population had an equal chance of being contacted because each student is required to have a unique “ncsu.edu” email.
Of those invited to participate, 712 undergraduates and 731 graduate students completed the survey, a total of 1,443, resulting in a 21% response rate. Assuming a 50-50 division in opinion calculated at a 95 percent confidence level, the margin of sampling error for the sample is (plus or minus) +/-2.5% for questions answered by the full sample. However, sampling error increases for each estimate when three or more choices were offered, for example, when fewer respondents were asked a version of a question, or when analyzing the opinions of sub-groups, such as when looking just at how female respondents answered a question.
The response rate for just the undergraduate student sample is 20%, and 21% for graduate students. Likewise, the margin of sampling error for describing the opinions of just undergraduates in this survey is +/-3.6%, while it is +/-3.4 for graduate students.
In addition to sampling error, other forms of non-sampling error occur in surveys, such as confusion about question wording or the order of questions, and non-response bias (low response rates), but these types of error are not precisely quantifiable.
We only apply post-stratification sample weights for student status when describing the opinions of “all” students. Undergraduates comprise 75% of the student population, but just 49% of our sample, so weighting is necessary to accurately describe the views of all students. There is a slight under-representation of males in our undergraduate sample, but not enough to warrant further weighting of that population it would not produce noticeable effects on any of the results.