The Staging Effect during House Sales

How to cite

Bhattacharya, Puja; Li, Sherry; Wang, Yu; Zheng, Xiang. 2025. "The Staging Effect during House Sales." AEA RCT Registry. AEARCTR-0016398. https://doi.org/10.1257/rct.16398-1.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Puja Bhattacharya* , Sherry Li , Yu Wang , Xiang Zheng

* Primary investigator

Registry ID
AEARCTR-0016398
DOI
https://doi.org/10.1257/rct.16398-1.0
First registered
Last updated
Status
In development
Trial dates
2025-07-22 – 2027-06-30
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/16398

Abstract

A home purchase is arguably the most important consumption decision for an average household in the United States. Increasingly, homes for sale are staged with furniture to help potential buyers imagine living in the house during in-person visits and online browsing. Since buyers do not acquire any of the staged furniture upon purchasing the home, classical economic theories predict that the home sale prices should not be affected by whether a home is staged or not. We test this conjecture by applying a machine learning algorithm to a unique empirical dataset combined between ZTRAX data and hand-collected data from another major house listing service. We find that houses with furniture sell 10% higher than those of similar quality but without furniture. This price premium is not explained by observable house characteristics, location, school districts, or the market conditions during transactions in our cross-sectional analyses. To complement the empirical study and further investigate the factors that could explain the price difference, we design and conduct an online experiment on Prolific to understand potential economic and behavioral mechanisms that may contribute to the higher valuations of furnished homes.

Experimental design

Our design is a randomized intervention embedded within a survey. Eligibility to participate in the survey includes: participant being between 25 and 75 years old, residing in the US, and personal income above $40,000. Participants are recruited through Prolific.

The experiment consists of three parts.

In Part 1, participants are shown images of three randomly selected homes that were listed for sale in the US (from a set of eight homes) with within-subject variation in the staging status of the homes - home is without furniture and decor (T1), home is with furniture but no decor (T2), and home is with furniture and decor (T3). Participants see five images for each home - the exterior, the kitchen, the living, dining and master bedroom. The latter three images of the homes are subject to the treatment variation. For each home, participants report their hypothetical willingness-to-pay, and also answer some questions about their perceptions of these homes.

In Part 2A, the participants are shown the same three homes as in Part 1, and guess the selling price and days on the market for each home. Their response is incentivized by rewarding them for an accurate prediction.

In Part 2B, all participants are randomly shown one version, either furnished or unfurnished, of the same bedroom from a home they have not seen in earlier parts of the experiment. Their task is to estimate the room’s dimensions (width and length). Accurate guesses are incentivized through a reward system.

In Part 3, the participants are shown two new homes, and have to hypothetically select one home they wish to live in. Participants are assigned randomly to one of three treatment arms - both homes are without furniture and decor (B1), the left home is without furniture and decor while the right home is with furniture and decor (B2), the left home is with furniture and decor while the right home is without furniture and decor (B3). When shown with furnishings, the two homes use the same set of furniture and decor.

Design details

Randomization method
Randomization done in Qualtrics
Randomization unit
Individual participant
Clusters
1500 participants
Observations
1500 participants
Arms
1500 total participants. Randomization done in Part 1, Part 2A, and Part 3. Parts 1 and 2A: within-subject intervention (1500 in T1, 1500 in T2, 1500 in T3) Part 3: between-subject intervention (500 in B1, 500 in B2, 500 in B3)
Primary outcomes
We have four outcome variables - 1. The willingness-to-pay for homes reported by participants 2. Participants' beliefs about the final selling price of the homes 3. Participants' beliefs about the number of days a home was on the market 4. Participants' preference over homes
Secondary outcomes
1. Participants’ estimates of the room’s dimensions; 2. Their perceptions about the homes.

Cited by

No citation data available for this registration yet.

Keywords