Cross-cultural differences in the social norm to work
How to cite
Alhumaid, Bodour; Howley, Peter; Najm, Sarah; Hetschko, Clemens. 2024. "Cross-cultural differences in the social norm to work." AEA RCT Registry. AEARCTR-0014245. https://doi.org/10.1257/rct.14245-1.0
Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.
* Primary investigator
Abstract
Research on how social norms influence Labour market behaviors and outcomes has flourished over the past few decades. For instance, it has been argued that there is a social norm to work which explains why unemployment negatively impacts workers’ mental health and wellbeing (e.g., Günther et al. 2024). The norm might also contribute to their job search efforts and (un)willingness to claim financial assistance they are eligible for.
Previous studies in this context have relied on indirect methods of eliciting this norm to work, for example, by using proxies such as the local unemployment rate (e.g., Clark 2003) or the local support for cuts in unemployment benefits (Stutzer and Lalive 2004). However, indirect proxies are, in principle, open to alternative interpretations. Moreover, it is to date largely unclear what exact behavior this supposed social norm to work prescribes (Chadi and Hetschko 2024). For example, does the norm require people to do market work or do household work or volunteering suffice to meet the norm? Another open question is whether the norm’s strength depends on individual characteristics, whereby obvious guesses are gender and the cultural background of a person. For instance, there is a lack of research on the topic in non-Western contexts. Answers to these questions will inform society about starting points for interventions alleviating the misery of the unemployed and bear implications for tailoring job search incentives and benefit policies.
The methodology of the project adopts the elicitation method for social norms proposed by Krupka and Weber (2013). In the process, incentivized survey experiments (factorial surveys) are used to causally identify the existence and strength of the social norm to work across a variety of contexts. Participants are presented with different scenarios of hypothetical choices, in our case, an unemployed person who accepts/rejects a job offer under certain circumstances (their gender, time use, race, culture), and are asked to indicate how socially appropriate they believe the majority of participants will deem the choice.
What is more, we are interested in the extent to which the provision of information can alter people’s assessment of what is socially appropriate and, hence, their perception of the social norm to work. To this end, we provide participants with information potentially altering their beliefs about the majority view. We inform randomly selected respondents that an increasing number of women are working outside the household, or that many unemployed people do not claim the financial assistance they are entitled, too.
Experimental design
In keeping with Krupka and Weber’s (2013) method of eliciting social norms, participants are randomly assigned different versions of a hypothetical scenario in which an unemployed decision taker named X accepts or rejects a job offer (first set of conditions). To test for the influence of individual characteristics, we randomly vary the gender (male/female) of the decision-taker (second set of conditions) and their time use during unemployment (third set: household work/leisure/volunteering). In the USA, we also randomly vary Person X’s race (fourth set: black/white).
The strength of the social norm is determined by asking respondents to indicate how socially appropriate they believe most participants deem X’s behavior under the various conditions assigned to them on a four-point scale (from totally inappropriate to totally appropriate). We will conduct this experiment across six countries, targeting a total of at most 3,250 participants: max. 1,000 from the USA and 450 each from Brazil, South Africa, the UK, Saudi Arabia, and India (dependent on the total cost and budget constraint, these numbers may be lower). In all countries, the factorial survey will randomly vary the choice scenario across 12 different treatment conditions (2*3*2). In the US, we include race (Black/White) as an additional condition, resulting in 24 different treatment conditions (2*3*2*2). To incentivize respondents to guess correctly, they enter a prize draw and possibly receive a financial reward (in the form of extra points) on top of the participation fee if they are able to guess the modal response correctly.
To test the influence of information provision on the perceived social norm to work, we also run a second RCT using randomized informational treatments, where we vary the information we convey to participants in three ways: more and more women are now working outside the home / many unemployed people do not claim welfare benefits even though they are eligible to receive these benefits / no treatment information.
The survey also includes questions relating to participants’ behavior, attitudes, perceptions, alongside other personal questions. As some of these personal questions may be deemed sensitive (e.g. income, health), we make sure to almost always include a ‘Prefer not to say’ option and have obtained ethical approval. In addition to that, we use the services of a well-known survey institution to collect and anonymize the data such that we do not handle any sensitive data ourselves.
Design details
Cited by
No citation data available for this registration yet.