A local news agency conducted a survey about unemployment by randomly dialing phone numbers until they had gathered responses from 1000 adults in their state. In the survey,
of those who responded said they were not currently employed. In reality, only
of the adults in the state were not currently employed at the time of the survey. Which of the following best explains the difference in the two percentages? (a) The difference is due to sampling variability. We shouldn’t expect the results of a random sample to match the truth about the population every time. (b) The difference is due to response bias. Adults who are employed are likely to lie and say that they are unemployed. (c) The difference is due to undercoverage bias. The survey included only adults and did not include teenagers who are eligible to work. (d) The difference is due to nonresponse bias. Adults who are employed are less likely to be available for the sample than adults who are unemployed. (e) The difference is due to voluntary response. Adults member of the sample.
The Correct Answer and Explanation is:
Correct answer: (d) The difference is due to nonresponse bias. Adults who are employed are less likely to be available for the sample than adults who are unemployed.
Explanation
In this scenario, the survey was conducted by randomly dialing phone numbers until 1000 adults responded. Although the selection method was random, the fact that the survey relied on who answered and completed the phone call introduces a nonresponse bias. This type of bias occurs when certain groups of people are systematically less likely to respond to a survey, resulting in a sample that is not representative of the population.
In particular, employed adults are more likely to be busy during the day, at work, commuting, or otherwise unavailable to answer a phone call and complete a survey. On the other hand, unemployed adults may be more available and willing to respond because they have more free time or may be more interested in discussing their employment status. As a result, the sample overrepresents unemployed individuals and underrepresents employed ones.
This leads to a distorted result where a larger percentage of respondents report being unemployed compared to the actual population rate. Even though the sampling method was random (random dialing), the final group of respondents was not random due to differential availability and willingness to participate, which is the essence of nonresponse bias.
Let’s examine the other choices briefly:
- (a) Sampling variability is always a factor, but it’s unlikely to fully explain a significant difference if the sample is large (like 1000) and systematically biased.
- (b) Response bias suggests dishonesty, but there’s no evidence that employed people lie about being unemployed.
- (c) Undercoverage bias refers to groups left out of the sampling frame; however, teenagers aren’t the focus here since the sample includes only adults.
- (e) Voluntary response would apply if people chose to opt into the survey on their own (e.g., through a public poll), but this was a random phone dialing method.
Thus, (d) nonresponse bias is the most accurate explanation.
