Skip to content
AP Statistics · Exploring One-Variable Data and Collecting Data

Survey bias

Survey bias appears in 7 reviewed AP Statistics questions and sits in Exploring One-Variable Data and Collecting Data. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

How Survey bias is tested

Example 1Medium
A newspaper runs an online poll asking: 'Do you support the city's controversial new transit plan? Yes or No.' Of the 2,300 people who responded, 78% said 'No.' What is the PRIMARY reason why this result should NOT be used to conclude that 78% of city residents oppose the transit plan?
  1. AThe question asks only for a yes/no answer, which introduces response bias by forcing residents with mixed views to choose a side.
  2. BThe sample size of 2,300 is too small to represent the city.
  3. CThe poll should have used a 5-point Likert scale instead of a yes/no format, which would reduce nonresponse bias.
  4. DVoluntary response bias: people with strong opinions are more likely to answer an online poll.Correct

Why D is right

An online poll with open participation suffers from voluntary response bias: people who respond are self-selected and tend to have stronger opinions (especially negative ones) than the general population. The 78% 'No' result is likely an overestimate of opposition in the full city population because strong opponents are more motivated to respond.
Why the other 3 choices miss
A
Incorrect because a yes/no format is a valid question structure. Response bias from question wording would apply if the question were leading, but here the question simply asks for a position. The bias comes from who responds, not from the answer format.
B
Incorrect because 2,300 is actually a large sample by statistical standards. The problem is not sample size but the self-selection of respondents, which creates bias regardless of sample size.
C
Incorrect because using a Likert scale would not address the fundamental problem of voluntary response bias. The issue is who chooses to respond, not the response format.

Studied alongside

Other concepts from Exploring One-Variable Data and Collecting Data.