A data science team is developing a survey to measure the prevalence of unregistered, off-the-books employment within a specific industry. The team anticipates that direct questioning will lead to significant underreporting due to social desirability bias and fear of legal repercussions. To obtain more accurate data while maintaining respondent anonymity, which survey generation technique should the team implement?
Distribute a voluntary response survey through industry forums.
Implement a list experiment (item count technique).
Employ a forced-choice Likert scale with an even number of points.
Utilize stratified sampling based on job roles within the industry.
The correct answer is to use a list experiment, also known as the item count technique. This indirect questioning method is specifically designed to elicit more truthful answers for sensitive topics by protecting respondent anonymity. In a list experiment, the sample is split into a control group and a treatment group. The control group is given a list of non-sensitive items and asked to report how many of the items are true for them, not which ones. The treatment group receives the same list plus the sensitive item. The prevalence of the sensitive behavior is then estimated by calculating the difference in the average number of items reported between the treatment and control groups. This approach reduces social desirability bias because respondents never have to directly admit to the sensitive behavior.
Using a Likert scale with an even number of points (a forced-choice scale) removes the neutral option but does not inherently solve social desirability bias; respondents might still choose the most socially acceptable non-neutral option. This type of bias is a known limitation of Likert scales.
Stratified sampling based on job role is a sampling strategy that ensures representation across different subgroups but does not address the measurement error caused by respondents giving untruthful answers to sensitive questions.
A voluntary response survey is a non-probability sampling method where individuals choose to participate. This method is highly susceptible to self-selection bias and is unlikely to attract individuals who wish to admit to illegal or socially undesirable behavior, thus worsening the underreporting problem.
Ask Bash
Bash is our AI bot, trained to help you pass your exam. AI Generated Content may display inaccurate information, always double-check anything important.
What is a list experiment, and how does it work?
Open an interactive chat with Bash
What is social desirability bias, and why does it impact surveys?
Open an interactive chat with Bash
Why is a Likert scale not effective for sensitive surveys?