While watching a well-known TV news channel the other day, a report came on about student reactions to the enhanced security procedures at U. S. airports. The correspondent indicated that students are split right down the middle, 50-50 on the use of full-body airport scanners. Fifty percent favor, fifty percent oppose. I was curious what data there were to support this contention. The correspondent stated that half the students he talked to were in favor, half opposed. He then proceeded to present the results of his sample, which appeared to be N = 2, one student in favor, one student opposed. But he went on, according to another poll, most Americans favor the full-body scanners. Now “most” is a most ambiguous word in my mind. By definition, most simply means the majority, but if asked to attach a number to the word “most,” I tend to think about 80%. So I concluded that about 80% of Americans favor the new scanning procedure. But just to be sure, I looked up the poll that was being cited here. ( http://www.washingtonpost.com/wp-srv/politics/polls/postpoll_11222010.html?hpid=topnews)
The poll indicated it was based on a random sample of 514 adults who were called either on their land line or cell phones. Now a random sample is one in which each member of a a population has an equal chance of being included in the sample. But I for one, have no chance of being included in this sample; I never respond to telephone polls, sometimes quite ungraciously. So the question arises, how many calls were made to obtain 514 responses? And how would those who declined to participate in this poll have responded?
The poll results indicate that if we consider only those people who “strongly support” the new body scanners, then 37% of the sample responded favorably. Twenty-seven percent were somewhat in favor, for a total of 64% responding that they support the scanning, either strongly or somewhat. But only 48% responded in a favorable way to the enhanced hand searches. And it also seems reasonable to expect that those who fly infrequently or not at all (53% of the respondents in the survey) may differ in their beliefs from those who fly frequently. We might also expect differences related to age of respondents, but we can’t tell from the results of the poll.
In her recent blog post, Bonnie included data collection methods as one of the core concerns for an introductory statistics course. To quote:
“Though I have said to my students more times that I can count, ‘the quality of our statistics is limited by the quality of our sample,’ I must admit to being a bit surprised that this was considered critical by others, especially since when I look at many undergraduate statistics textbooks, data collection methods are barely mentioned.”
The two examples given above provide excellent support for Bonnie’s contention that students should be taught to carefully evaluate the quality of the data on which statistics are based. Who were the subjects and how were they selected? What questions were asked and what responses allowed? And what inferences can be made from the results. It might prove to be an interesting class exercise to have students find media reports of current polls and then actually access the poll to see who the respondents were, what questions were asked, and the results obtained.