Skip to content
No Vape

What can a list experiment estimate about vaping?

Read the difference between group item counts while keeping individual answers unknown.

No Vape · Updated

Disposable vape resting on concrete
Illustrative image · Markus Winkler / Unsplash

Read what respondents actually provided

Blair and Imai describe the standard design as a comparison of randomly assigned forms. The added item is the sensitive topic of interest. People give a total count for the list they receive, so the ordinary reported response does not state which individual items applied.

For an invented vaping example, the short list asks about three activities over a stated period. The longer list adds a defined vaping behavior. Check the paper for the exact wording and time window if reading a real application. “Vaped at least once this week” would answer a different question from frequency of use or an intention to quit.

Political Analysis: Statistical Analysis of List Experiments

Keep the difference at the group level

Imagine the fictional short-list group reports an average of 1.2 items. The longer-list group reports an average of 1.5. The difference is 0.3 items per respondent. Under the standard assumptions, that corresponds to an estimated 30 percent affirming the added item for the group in the example. No actual vaping survey supplied these numbers.

One person reporting a count of two from the longer list may have affirmed the added item or two ordinary items. A single count generally leaves that distinction unresolved. The arithmetic does not give a reader grounds to assign a particular vaping answer to that person.

Check the assumptions about the shared items

The methods paper explains that adding the sensitive item should not change the total of affirmative responses to the original items. Suppose seeing the vaping item makes someone reinterpret another item. Then the difference in averages could reflect more than the extra behavior. The standard analysis also relies on respondents following the procedure and reporting the sensitive part truthfully.

Find how the forms were assigned and whether the authors discuss checks, uncertainty and problems when a person’s count falls at an extreme. The label “indirect” does not guarantee that a real design protects every person’s answer or eliminates reporting problems.

Political Analysis: Statistical Analysis of List Experiments

State the inferred result alongside the method

A useful reading note gives the two list versions, the average difference, the uncertainty and the exact behavior named by the added item. Leave any unreported details open. This method example explains a population-level reading question, without claiming that a particular vaping survey used the design.

Questions about this step

Why can’t I read the long-list average as a vaping rate?

It includes affirmative counts for all items in that list. The standard comparison uses the difference against the short-list group under the stated assumptions.

Sources and scope

Sources inform the guidance. The worksheets and examples are No Vape editorial suggestions, not tested treatments.

  1. Political Analysis — Statistical Analysis of List Experiments

Written with AI assistance and checked against the cited sources. No clinical review is claimed. This guide covers practical planning, not individual medical treatment. How we research and write