Information Spillovers within Couples: Evidence from a Sequential Survey of Spouses

How to cite

Zafar, Basit; Kosar, Gizem; Delavande, Adeline. 2025. "Information Spillovers within Couples: Evidence from a Sequential Survey of Spouses." AEA RCT Registry. AEARCTR-0016381. https://doi.org/10.1257/rct.16381-1.0

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

Basit Zafar* , Gizem Kosar , Adeline Delavande

* Primary investigator

Registry ID
AEARCTR-0016381
DOI
https://doi.org/10.1257/rct.16381-1.0
First registered
Last updated
Status
Completed
Trial dates
2022-01-19 – 2022-07-15
Countries
United States of America ()
JEL codes
D83, H31, J26
Registry record
http://www.socialscienceregistry.org/trials/16381

Abstract

To understand the extent and drivers of information flow within couples, we conduct an online survey of 2,200 middle-aged married couples in the US. Our focus is on expectations about Social Security benefits. We first investigate the correlation between partners’ beliefs about a given spouse’s Social Security benefits. We analyze whether the imperfect alignment is systematically associated with couple-specific characteristics. To establish causal evidence on information spillovers, we implement a randomized information experiment paired with a sequential survey design, where the index spouse receives targeted information, and the other is surveyed a few days later. Using detailed survey data on measures of communication frictions, cognitive barriers, and the value of information, we identify key drivers of information flow.

Experimental design

Our design is a randomized information intervention embedded within a survey. Eligibility to participate in the survey includes: index spouse being between 38 and 62 years old, being married, not retired, not receiving Social Security benefits, reporting that spouse is not receiving Social Security benefits, and willing to share their spouse’s email address. Index spouses who share an email address that is not different than their own for their spouses are not admitted to the survey. Secondary spouses receive the survey link after the index spouse finishes the survey.

The randomization is done at the couple level. Couples are divided into two groups:
i) Control (22% of the couples)
ii) Treatment (78% of the couples)

For secondary spouses, the treatment groups is divided into two groups:
a) Treatment 1 (39% of the secondary spouses)
b) Treatment 2 (39% of the secondary spouses)

T1 provides information about predicted SS benefits (based on the SS Quick Calculator) for own self and for one's spouse. This is done in stages for the index spouse, and done at the same time for the secondary spouse.
T2 is the same as T1, except that the secondary spouse is also given information about the reported beliefs of the index spouse about the age at which they expect to claim benefits.

Design details

Randomization method
Randomization is done randomly by the computer by the survey vendor, NielsenIQ.
Randomization unit
Randomization is done at the couple level.
Clusters
Not clustered
Observations
2219 couples
Arms
492 control couples 1727 treatment couples For secondary spouses: 856 spouses for Treatment 1 871 spouses for Treatment 2
Primary outcomes
i. Beliefs about own Social Security benefits at age 67 ii. Beliefs about spouse's Social Security benefits at age 67 iii. Beliefs about own Social Security benefits claiming age iv. Beliefs about spouse's Social Security benefits claiming age These are used to construct different variables of interest, such as perception gaps (explained below).
Secondary outcomes
i. Perception gaps about own Social Security benefits at age 67 = Predicted benefits by the Social Security calculator - Beliefs about own Social Security benefits age age 67 ii. Perception gaps about spouse's Social Security benefits at age 67 = Predicted benefits by the Social Security calculator - Beliefs about spouse's Social Security benefits age age 67 iii. Distance between spouses’ reported expected claiming ages for the same spouse

Cited by

No citation data available for this registration yet.

Keywords