2026
By Zilin Lin and Alexandra Schwinges
Search engines shape how users encounter information, yet little is known about how biased interface components affect users. Existing studies often focus on isolated components like ranked results and rely on simplified designs, limiting ecological validity. This project develops the Search Experiment Builder (SEB), an open-source platform for manipulating key interface components, including autocomplete suggestions, AI summaries, ranked results, and recommended keywords in controlled yet realistic search experiments. SEB is showcased through a pre-registered experiment on out-group bias in search results. The project contributes a reusable tool and new evidence on biased search interaction.