When should I use a Double List Experiment?

Double List Experiments (DLE) are recommended to reduce variance by using all respondents for estimation, though this skill also guides you through the necessary carryover diagnostic tests required for DLEs.

Does it support the 'list' R package?

Absolutely. It is designed to work alongside the 'list' package in R, providing guidance on using functions like ict.test(), ictreg(), and ict.hausman.test().

What is a list experiment?

A list experiment, or item count technique (ICT), is a survey method designed to measure sensitive attitudes or behaviors by providing respondents with plausible deniability, reducing social desirability bias.

How does this skill help with estimator selection?

The skill provides a selection framework for choosing between Difference-in-Means, Nonlinear Least Squares (NLSreg), and Maximum Likelihood (MLreg) based on your survey's susceptibility to measurement error and inattentive responding.

Can this skill diagnose survey failures?

Yes, it includes instructions for testing identifying assumptions like 'No Design Effect' (NDE) and checking for ceiling or floor effects that might bias your results.

List Experiment Designer

Name: List Experiment Designer
Author: scdenney

byscdenney

•

データサイエンスとML

Guides the design, implementation, and statistical analysis of list experiments to measure sensitive behaviors accurately.

This skill provides a comprehensive framework for conducting list experiments (Item Count Technique), a survey method used to elicit honest responses on sensitive topics. It assists researchers through the entire pipeline—from evaluating if a list experiment is warranted based on social reference theory to selecting control items, choosing design variants (single, double, or placebo), and performing rigorous diagnostic tests. By codifying standards from methodological texts, it helps users mitigate ceiling and floor effects, test for design-effect violations, and select the most robust estimators, such as NLSreg or the combined estimator, to ensure high-quality statistical inference.

主な機能

01Implements diagnostic workflows for No-Design-Effect and floor/ceiling assumption testing

0215 GitHub stars

03Supports advanced design variants including Double List (DLE) and Placebo-item designs

04Evaluates sensitivity bias and precision trade-offs to determine if a list experiment is warranted

05Guides selection between Difference-in-Means, NLSreg, and MLreg estimators

06Provides constraints and prevalence targets for designing effective control item lists

ユースケース

01Optimizing statistical power and estimator choice for sensitive question analysis in social science research

02Performing diagnostic tests on existing survey data to identify mechanical inflation or artificial deflation

03Designing survey instruments to measure socially stigmatized behaviors like prejudice or illegal activities

主な機能

01Implements diagnostic workflows for No-Design-Effect and floor/ceiling assumption testing

0215 GitHub stars

03Supports advanced design variants including Double List (DLE) and Placebo-item designs

04Evaluates sensitivity bias and precision trade-offs to determine if a list experiment is warranted

05Guides selection between Difference-in-Means, NLSreg, and MLreg estimators

06Provides constraints and prevalence targets for designing effective control item lists

ユースケース

01Optimizing statistical power and estimator choice for sensitive question analysis in social science research

02Performing diagnostic tests on existing survey data to identify mechanical inflation or artificial deflation

03Designing survey instruments to measure socially stigmatized behaviors like prejudice or illegal activities