QuestionPunkQuestionPunk
무료로 시작하기
모든 템플릿

Sensitive Topic List Experiment (Item Count)

Measure behaviors people won't admit directly: the list experiment (item count technique) asks only HOW MANY statements apply — never which — so individual answers stay genuinely deniable while group comparisons reveal the true rate. The native question type randomizes control and treatment lists for you.

이 템플릿 사용하기템플릿 더 둘러보기

샘플 질문

템플릿에 포함된 내용을 미리 확인해 보세요. 모든 질문은 설문 공개 전에 자유롭게 수정할 수 있습니다.

질문 6개 · 약 5분
Q01
메시지

This study uses a technique designed for honest answers on sensitive topics: you'll see a short list of statements and tell us only HOW MANY apply to you — never which ones. Your responses are completely confidential and anonymized. That means no answer you give can reveal anything specific about you.

Q02
질문필수

How many of the following statements apply to you? Count them privately, then enter only the number.

Q03
의견 척도필수

How comfortable did you feel answering honestly with this counting format?

척도: 1 – 5
최소:Still felt exposed최대:Completely safe
Q04
객관식필수

In general, how sensitive do you consider this topic among your peers?

  • Not sensitive — people discuss it openly
  • Somewhat sensitive
  • Very sensitive — rarely discussed honestly
Q05
AI 인터뷰

WITHOUT ever asking whether the sensitive statement applied to them, explore the topic's social context: why people in their environment might underreport this behavior in normal surveys, what social or professional consequences drive that, and what conditions (anonymity guarantees, framing, who's asking) make honest answers more likely. Keep the tone academic and never probe their personal count.

Q06
메시지

Thank you. Because different participants received slightly different lists, comparing group averages estimates how common the sensitive behavior really is — with no individual ever identifiable.

포함된 기능

  • AI 후속 질문

    정형화된 설문이 놓치는 세부 내용을, 주관식 답변에 맞춰 AI가 심층 질문으로 끌어냅니다.

  • 주의력 확인 장치

    성의 없는 답변과 저품질 응답자를 걸러내는 내장 안전장치입니다.

  • AI가 작성한 문안

    문구, 질문 순서, 분기 로직까지 AI가 연구 목표에 맞춰 작성합니다.

  • 자동 리포트

    응답이 모이면 주요 주제, 인용문, 이해하기 쉬운 요약이 자동으로 작성됩니다.

다른 서비스와 비교

다른 설문 도구의 가장 유사한 템플릿을 검토했습니다. 그 도구들이 잘하는 점과, 이 템플릿이 한발 더 나아가는 지점을 정리했습니다.

이 템플릿을 선택하는 이유

  • A native list-experiment question type randomizes control and treatment lists automatically — the method's hardest part, handled
  • Individual answers stay genuinely deniable: respondents only ever report a count, never which items
  • A comfort check validates that the format actually made honest answering feel safe
  • The AI interview explores the topic's social context without ever probing any individual's answer

Gradient Metrics

List Experiments

Applied methodology guide (not a drop-in survey template) explaining list experiments as indirect measurement of private opinion via the item-count technique: respondents report how many of a list they agree with, control vs. treatment lists differ by one sensitive item, and prevalence is the difference in means. Cites a real 19,000+ response deployment (Social Pressure Index with Populace). No competitor here ships this as a native, self-serve question type.

잘하는 점

  • Clear, correct explanation of the control-vs-treatment design and difference-in-means estimation
  • Grounds the method in a real large-sample deployment (19,000+ responses, Social Pressure Index)
  • Frames the practical use case: measuring the gap between public and private opinion on sensitive topics
  • Emphasizes the privacy guarantee that makes honest answers possible

아쉬운 점

  • It is a blog/methodology explainer, not a usable template or a built-in question type a researcher can drop into a survey
  • No tooling to auto-randomize respondents into control/treatment arms and enforce balanced allocation
  • No built-in estimator/report that computes prevalence and confidence intervals from collected data
  • No guardrails against ceiling/floor effects (list design) surfaced for a non-methodologist user

SensitiveQuestions.org (R 'list' package)

Statistical Methods for the Item Count Technique and List Experiment (R package 'list')

The canonical academic toolkit for list-experiment analysis: an R package implementing multivariate/random-effects/Bayesian MCMC regression, joint modeling, combined list+direct-question estimators, and statistical tests to detect list-experiment failure. Authoritative on analysis, but it is code for researchers post-collection, with no fielding UI, randomization, or respondent experience.

잘하는 점

  • Comprehensive, peer-reviewed estimators (multivariate, random-effects, Bayesian MCMC hierarchical regression)
  • Supports advanced designs: multiple sensitive items, list experiments as predictors, combined list+direct estimates
  • Includes diagnostics and placebo tests to detect list-experiment failure
  • Grounded in six methods papers (2011-2016), the field standard for analysis

아쉬운 점

  • Analysis-only R code; provides no survey fielding, randomization, or respondent-facing UI
  • Requires statistical programming expertise, out of reach for a typical survey author
  • No integration with data collection: the researcher must field the experiment elsewhere and export data
  • No auto-generated plain-language report; outputs are statistical objects, not decision-ready summaries

설문을 공개할 준비가 되셨나요?

이 템플릿을 편집기에서 열어 보세요. 첫 응답자가 보기 전에 모든 부분을 원하는 대로 바꿀 수 있습니다.

Use this template무료로 시작하기