Bias in a Survey: Leading Questions and Unfair Samples

0:00A survey asks people questions to learn what a group thinks. It is biased if the way it is run pushes the results one way.

0:09Bias can hide in two places: the question you ask, and the people you ask.

0:16Here is a question: do you agree that the long, boring assembly should be shorter?

0:22Long and boring tell you what to think, and do you agree invites a yes.

0:28A question that pushes toward one answer is called a leading question.

0:34Does asking more people fix it? Suppose 16 of 20 students say yes. 16 divided by 20 is 80 percent.

0:43Ask 200 students the same question: if 160 say yes, 160 divided by 200 is still 80 percent.

0:54Ten times the answers, the same push. More answers do not fix a leading question.

1:01The fix is a neutral question that hints at no answer: how long should assemblies be?

1:08A second survey asks: should the school spend more on sports?

1:13The population is the whole group you want to learn about: every student in the school.

1:19The sample is the people you actually ask: here, only the basketball team.

1:26Players are more likely to want sports money, so the sample favours one group. This is sampling bias.

1:34The fix is a random sample: every student has the same chance of being picked.

1:40A fair survey asks a neutral question to a random sample.