Goal: to demonstrate how bias affects survey results
Time: 10 minutes
In this exercise, students are given handouts that show numbered counters scattered on a checkerboard surface. Half of the students sum the counters lying on black squares, while the rest add up those lying on white squares. As some of the counters lie across two different squares, it will be up to the students to decide what colour they fall onto. This exercise works best if the students do no know that someone else is counting the opposite colour.
"There is a total score of 60, with 30 lying on each colour. But did the results reflect that? If not, why not? Would the result be different if you ran the exercise again?"
Getting rid of unwanted bias is one of the most important issues in opinion surveys (and science in general!), but also one of the hardest to solve. There is a seemingly never-ending list of things that can unduly influence the result of a survey. Who is asked, how they are selected, the wording of the questions, how the data is collected; all of these demand careful consideration by the pollster and consequently by any journalist reporting the results of the survey.