Measuring subjective probabilities about future income: the effect of question modes on the use of focal responses and moments of the predictive distribution

Last registered on January 13, 2025

Pre-Trial

Trial Information

General Information

Title
Measuring subjective probabilities about future income: the effect of question modes on the use of focal responses and moments of the predictive distribution
RCT ID
AEARCTR-0015009
Initial registration date
January 13, 2025

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
January 13, 2025, 2:13 PM EST

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Affiliation
Banco de España

Other Primary Investigator(s)

PI Affiliation
Banco de España
PI Affiliation
Banco de España
PI Affiliation
CUNEF

Additional Trial Information

Status
In development
Start date
2025-01-13
End date
2025-07-31
Secondary IDs
Prior work
This trial does not extend or rely on any prior RCTs.
Abstract
The ability to assess probabilities of future outcomes is an essential component of individuals’ decision-making under risk. A growing number of surveys ask people to assess their subjective probabilities for future events, and there is ample evidence that these beliefs matter for individuals’ choices. At the same time, different surveys employ different formats when asking for beliefs, and answers to these expectations questions show inconsistencies and measurement error. The objective of this project is to study how different formats relate to the measures of beliefs we derive from them. We implement an online survey of 4,400 participants to test how the format of income-expectation questions relates to measures of coherence, rounding, and estimated moments of the underlying beliefs. The survey randomizes the format of the income-expectation question, between a baseline format based on distributing points in a pre-set scale of income growth, a second and similar format distributing more points in more intervals, and a third min-max format based on distributing points among individual-specific intervals of income levels.
External Link(s)

Registration Citation

Citation
Basso, Henrique et al. 2025. "Measuring subjective probabilities about future income: the effect of question modes on the use of focal responses and moments of the predictive distribution." AEA RCT Registry. January 13. https://doi.org/10.1257/rct.15009-1.0
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
We will implement an online survey that randomizes participants into different formats of the income-expectation questions. The survey will be completed by 4,400 participants aged 25-64, representative of the population living in Spain. The goal is to measure how different expectations formats relate to coherence, rounding and estimated moments of the underlying beliefs.
Intervention Start Date
2025-01-13
Intervention End Date
2025-07-31

Primary Outcomes

Primary Outcomes (end points)
A battery of indicators related to the coherence, rounding, mean, variance, and asymmetry of responses to the income expectations questions.
Primary Outcomes (explanation)
For coherence, we consider bunching, proportion of intervals used per individual, and use of non-adjacent intervals. See Bover (2015).

For rounding, we will measure the proportion of individuals rounding probabilities (for example, to the nearest, 5, 10 or 25 percent). See Attanasio and Augsburg (2016).

Individual means and standard deviations of beliefs will be computed following Arellano et al. (2022a), which assumes a normal distribution, or discrete beliefs as in Pistaferri (2001). As a measure of income risk, we will compute the coefficient of variation proposed by Arellano et al (2022b).

For asymmetry measures we will turn to Basso et al. (2023).

References:
- Arellano, M., O. Attanasio, B. Augsburg, and S. Crossman (2022a). Estimating flexible income processes from subjective expectations data: evidence from colombia and india. Technical report, CEMFI.
- Arellano, Manuel, Stéphane Bonhomme, Micole De Vera, Laura Hospido, and Siqi Wei (2022b). Income Risk Inequality: Evidence from Spanish Administrative Records. Quantitative Economics, 13, .no 4, 1747-1801.
- Attanasio, O. and B. Augsburg (2016). Subjective expectations and income processes in rural india. Economica 83 (331), 416–442.
- Basso, H. S., O. Bover, J. Galvez, and L. Hospido (2023). Income uncertainty and nonlinear dynamics: A subjective expectations framework. Technical report, mimeo.
- Bover, O. (2015). Measuring expectations from household surveys: new results on subjective probabilities of future house prices. SERIEs 6 (4), 361–405.
- Pistaferri, L. (2001). Superior information, income shocks, and the permanent income hypothesis. Review of Economics and Statistics 83 (3), 465–476

Secondary Outcomes

Secondary Outcomes (end points)
As a secondary outcome we will look at the correlation between income expectations and consumption expectations and plans, conditional on observed characteristics, including current income.
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
There are 3 experimental arms, each with a different format of the question asking for income expectations in the following next 12 months. The first arm corresponds to asking for expectations of income growth, where individuals are asked to distribute 10 points among 5 different intervals. The second arm is similar, asking for expectations of income growth too, but where individuals are asked to distribute 100 points among 6 different intervals. The third arm corresponds to asking for income expectations in levels using the min-max format. That is, individuals are first asked the minimum and maximum expected income, and then they are asked to distribute 100 points among 4 equally sized intervals between the minimum and the maximum they reported.

Our target population corresponds to individuals living in Spain, between 25- and 64-year-old. We aim for 4,400 completed interviews, distributed evenly among the three experimental arms. This will be an online survey, for a convenience panel of the partner organization, who will randomize the question format, ensuring representative quotas in terms of gender, age group (25 to 34, 35 to 49, and 50 to 64), education level (primary, secondary and tertiary), and labor status (employed, unemployed, other).
Experimental Design Details
Not available
Randomization Method
After entering the online survey and providing some basic information related to the quotas, individuals are randomized by the provider system into one of the experimental arms.
Randomization Unit
individuals
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
4,400 individuals (same as number of observations)
Sample size: planned number of observations
4,400 individuals
Sample size (or number of clusters) by treatment arms
around 1,450 individuals per treatment arm
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB

Institutional Review Boards (IRBs)

IRB Name
Research Ethics Committee at Centro de Estudios Monetarios y Financieros (CEMFI)
IRB Approval Date
2024-12-09
IRB Approval Number
N/A