conjoint-experiment

To estimate AMCEs from conjoint survey experiment data, use a linear probability model with dummy variables and standard errors clustered by respondent to account for within-respondent correlation.

Your dataset needs a long format with respondent_id, task_id, profile_id, a binary outcome like chosen, and attribute columns encoded as categorical levels for dummy-variable OLS estimation.

Yes, you can analyze subgroup heterogeneity and interaction effects in a conjoint experiment by computing marginal means and re-estimating AMCEs by subgroups like party or ideology to quantify preference differences.

You can design randomized conjoint survey profiles for political science research by generating full or partial profile sets for tasks and respondents so each attribute level is comparably varied across tasks.

Conjoint analysis with clustered standard errors works for measuring candidate evaluation, policy trade-offs, immigration preferences, and partisan asymmetries by applying randomized conjoint tasks to respondent profile-choice data.

Clustered standard errors are used when estimating marginal means in survey experiments to respect within-respondent correlation, ensuring accurate uncertainty estimates from multiple profile choices made by the same individual.