What does this research mean for the field?

The BRIDGE method, which leverages generative AI embeddings, effectively recovers true causal parameters and controls for confounding when using many unaltered real-world textual descriptions as experimental stimuli. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

To introduce and validate BRIDGE, a method using generative AI embeddings to incorporate vast, unaltered real-world textual stimuli into behavioral experiments while preserving causal inference.

September 8, 2026Journal of Marketing Research

EXPRESS: Behavioral Research Through Interpretable, Dimensionality-reduced Generative AI Embeddings (BRIDGE): A Method to Incorporate Real-World Stimuli in Consumer Experiments

View Full Paper

Ask AI

Bookmark

Authors

AMAnirban Mukherjee HCHannah H. Chang SGSachin Gupta

Discussion

Member takes

Overview

Experimental validation demonstrates causal parameter recovery using generative AI embeddings across diverse consumer choices, indicating enhanced realism in behavioral research.

Key Points

To introduce and validate BRIDGE, a method using generative AI embeddings to incorporate vast, unaltered real-world textual stimuli into behavioral experiments while preserving causal inference.
Developed an analytical framework using generative AI embeddings to construct low-dimensional, interpretable representations of focal constructs alongside statistical controls for non-focal nuisance variations.
Evaluated the method via Monte Carlo simulations, two coffee certification experiments, and a validation choice experiment with 1,000 participants evaluating approximately 50,000 unique descriptions sampled from 120,000 texts.
Simulations and laboratory evaluations confirmed that BRIDGE consistently recovers true parameters even in the presence of unobserved nuisance variations and complex confounding.
The large-scale choice experiment demonstrated that exposure to entirely incidental initial product descriptions significantly shaped participants' subsequent consumer preferences.

Cite This Study

Mukherjee et al. (2026) studied this question.

synapsesocial.com/papers/6a9fd77658e84d0ff5b46339 https://doi.org/10.1177/00222437261484068

View Full Paper

Ask AI

Bookmark

September 8, 2026Journal of Marketing Research

EXPRESS: Behavioral Research Through Interpretable, Dimensionality-reduced Generative AI Embeddings (BRIDGE): A Method to Incorporate Real-World Stimuli in Consumer Experiments

View Full Paper

Ask AI

Bookmark

Authors

AMAnirban Mukherjee HCHannah H. Chang SGSachin Gupta

Discussion

Member takes

Overview

Experimental validation demonstrates causal parameter recovery using generative AI embeddings across diverse consumer choices, indicating enhanced realism in behavioral research.

Key Points

To introduce and validate BRIDGE, a method using generative AI embeddings to incorporate vast, unaltered real-world textual stimuli into behavioral experiments while preserving causal inference.
Developed an analytical framework using generative AI embeddings to construct low-dimensional, interpretable representations of focal constructs alongside statistical controls for non-focal nuisance variations.
Evaluated the method via Monte Carlo simulations, two coffee certification experiments, and a validation choice experiment with 1,000 participants evaluating approximately 50,000 unique descriptions sampled from 120,000 texts.
Simulations and laboratory evaluations confirmed that BRIDGE consistently recovers true parameters even in the presence of unobserved nuisance variations and complex confounding.
The large-scale choice experiment demonstrated that exposure to entirely incidental initial product descriptions significantly shaped participants' subsequent consumer preferences.

Cite This Study

Mukherjee et al. (2026) studied this question.

synapsesocial.com/papers/6a9fd77658e84d0ff5b46339 https://doi.org/10.1177/00222437261484068

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

View Full Paper

Ask AI

Bookmark