Social norms, parenting behaviors and child aspirations: Two survey experiments in Ghana

How to cite

Aurino, Elisabetta; Thomas, Katherina; Emenalo, Chukwunonye Obi-Ogulo; Wolf, Sharon. 2023. "Social norms, parenting behaviors and child aspirations: Two survey experiments in Ghana." AEA RCT Registry. AEARCTR-0012461. https://doi.org/10.1257/rct.12461-2.0

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

Elisabetta Aurino* , Katherina Thomas , Chukwunonye Obi-Ogulo Emenalo , Sharon Wolf

* Primary investigator

Registry ID
AEARCTR-0012461
DOI
https://doi.org/10.1257/rct.12461-2.0
First registered
Last updated
Status
In development
Trial dates
2023-11-13 – 2026-07-31
Countries
Ghana (Greater Accra, Central, Eastern regions)
Registry record
http://www.socialscienceregistry.org/trials/12461

Abstract

This study aims to understand the influence of social norms on survey responses, focusing on parenting behaviors as reported by a sample of ~2,500 parents, and the educational and job aspirations of their adolescent children in Ghana. This sample is part of a larger intervention study (LEAD, Aurino et al. 2024).
We do so through two survey experiments, which we aim to investigate separately. With regards to parents, we test if random prior exposure to a survey module reminding them about social image around culturally sensitive parenting behaviors leads them to report differently on their own disciplinary practices towards their children. We assess whether there is an effect of eliciting social pressure norms on parents reports of their parenting practices, in a way that aligns with their perceived social norms in the community around these behaviors. We test heterogeneity by examining differences in effects by parental gender, socio-economic status (SES), baseline social desirability bias score, and enumerator gender.
For adolescents, we assess if random prior exposure to a survey module to measure child gender attitudes leads to differential reporting on their educational and job aspirations. In doing so, we test whether making gender norms more salient leads to lower reported aspirations and/or expectations for girls, consistent with the ‘stereotype threat’ literature, and boost such aspirations for boys, consistent with the ‘stereotype boost’ hypothesis. Further, we assess whether such effects are different for younger and older adolescents; parental SES, gender bias; and whether they vary by the gender of the enumerator.

Intervention

The intervention is the order in which adolescents receive two modules within a larger survey: (1) aspirations and expectations followed by gender bias scale (comparison), and (2) gender bias scale followed by educational aspirations and expectations (intervention). Caregivers also receive two modules within a larger survey: (1) parenting questions followed by social pressure module (comparison), and (2) social pressure module followed by parenting questions (intervention).

Update April 2026: we are replicating the experiment ONLY for adolescents in a new survey wave (2026), when adolescents are now aged 15 years on average to explore heterogeneity by age in the effects of priming gender norms on aspirations.

Experimental design

Within-survey randomization

For our estimates, we are going to compare outcomes of treated adolescents and caregivers with the control group’s outcomes to estimate treatment effects. We will estimate the following ordinary least squares regressions:
𝑌(i) = 𝛽0 + 𝛽1 𝑡𝑟𝑒𝑎𝑡𝑒𝑑 + 𝛽2X(i) + 𝜀!
Where:
𝑌(i) is the outcome of individual i, treated is the treatment status of individual i for the survey experiment, X is a vector capturing the
randomization strata within the original QP4G study and the adolescent intervention, and ε(i) is an individual error term. β1
captures our treatment effect.
We will also explore heterogeneity by selected axes, including:
• Adolescent, parent, and enumerator gender
• A measure of socioeconomic status (e.g. parent education or household poverty)
• Gender bias: Parental gender bias is measured by asking if parents agree or disagree
with the following statement: It is important that sons have more education than
daughters.
• Parental social desirability bias: as measured by Crowne and Marlowe (1960); in 2026, we have also measured adolescent social desirability bias

Design details

Randomization method
Within-survey randomization done via Stata dofile
Randomization unit
Adolescent/caregiver (2023-24 round) Adolescent only (2026 round)
Clusters
~2,500 adolescents, ~2,500 parents. The final number may vary based on how many parents and adolescents we will recruit.
Observations
~2,500 adolescents, ~2,500 parents. The final number may vary based on how many parents and adolescents we will recruit.
Arms
~1,250 adolescents comparison, ~1,250 parent comparison, ~1,250 adolescents treatment, ~1,250 parent treatment.
Primary outcomes
Educational and career aspirations and expectations for adolescents, self-reported disciplinary practices for parents.
Secondary outcomes
As exploratory outcomes, in the April 2026 re-run of the experiment, we also added questions on expected ages to: (i) start working/supporting family; (ii) leave full-time education; (iii) get married and start living with a partner; (iv) have a child. Questions are also sourced from Young Lives (Favara et al., 2021).

Cited by

No citation data available for this registration yet.

Keywords