Narrative Willingness to Pay

How to cite

Albertazzi, Andrea; Stringhi, Alessandro; Stimolo, Marco; Pin, Paolo. 2025. "Narrative Willingness to Pay." AEA RCT Registry. AEARCTR-0017081. https://doi.org/10.1257/rct.17081-1.0

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

Andrea Albertazzi* , Alessandro Stringhi , Marco Stimolo , Paolo Pin

* Primary investigator

Registry ID
AEARCTR-0017081
DOI
https://doi.org/10.1257/rct.17081-1.0
First registered
Last updated
Status
In development
Trial dates
2025-11-03 – 2026-04-09
Countries
Italy ()
JEL codes
C91; D81; D83; D91
Registry record
http://www.socialscienceregistry.org/trials/17081

Abstract

This study extends another ongoing project pre-registered here: https://doi.org/10.1257/rct.15672-1.0
We build on the framework of the existing project with the aim of exploring how narratives affect the monetary value people attach to nondiagnostic information.
We will replicate a similar Bayesian problem in a controlled experiment, where individuals face uncertainty and use new information to update their beliefs about the true state of the world. New information does not come for free, but signals can be purchased by the participant before stating their belief.

We will randomly assign participants to one of four conditions:
1) C_No_Info: participants engage with an abstract inference problem where acquired signals can only be diagnostic or missing.
2) C_Yes_Info: participants engage with an abstract inference problem where acquired signals can be diagnostic and nondiagnostic.

In the remaining treatments, participants will instead receive a fictional story, framed around identifying a culprit, which mirrors the same uncertainty and signals as the baseline but embeds them in a more emotionally and contextually rich framework.
3) N_No_Info: participants engage with the story, and acquired signals can only be diagnostic or missing.
4) N_Yes_Info: participants engage with the story, and acquired signals can be diagnostic and nondiagnostic.

By comparing differences across treatments, we assess whether the presence of a narrative increases the monetary value participants assign to nondiagnostic information. We expect higher valuations because, in our previous experiment, participants used nondiagnostic information to update their beliefs. Our goal is to show that subjects not only (mis)use nondiagnostic signals but also regard them as informative.

Intervention

We will vary two treatment variables in a 2×2 between-subjects design:
1) Type of context: participants are exposed to either an abstract or a narrative (fictional) story.
2) Signal technology: nondiagnostic signals are either present or absent (i.e., participants see a blank line instead of the nondiagnostic signal).

Experimental design

In all treatments, participants first read an initial ‘story’. In C_No_Info and C_Yes_Info, the story is presented in an abstract inference frame; in N_No_Info and N_Yes_Info, it is framed as a narrative. All stories, regardless of treatment, involve two mutually exclusive alternatives.

Participants state a prior belief about which alternative is correct.

Next, for each story they play 5 rounds in which they state the price at which they are willing to purchase three new signals. Prices are elicited with an incentivised Becker–DeGroot–Marschak (BDM) mechanism. Depending on treatment, signals can favour one of the two alternatives or be nondiagnostic/missing (i.e., no signal is shown). The ex-ante probabilities of each signal are common knowledge. After each BDM elicitation, participants report a posterior belief.

Each participant completes three different stories (15 rounds in total). Participants are paid for each belief elicitation.

Design details

Randomization method
At the beginning of each experimental session, the computer will randomly assign each participant to one of the four treatments.
Randomization unit
The randomization of the treatment is at the individual level. The order of the stories is randomized at the session level.
Clusters
Each experimental session constitutes one cluster for the randomization of the story order. The total number of clusters corresponds to the number of planned sessions, 24.
Observations
We plan to run laboratory sessions until we recruit at least 46 participants per experimental condition. Therefore, we will recruit at least 184 participants in total.
Arms
We plan to run laboratory sessions until we recruit at least 46 participants per experimental condition. Therefore, we will recruit at least 184 participants in total.
Primary outcomes
Our primary outcome is the price participants are willing to pay to receive new information, elicited via a Becker–DeGroot–Marschak mechanism (BDM).

Cited by

No citation data available for this registration yet.

Keywords