How language affects economic behavior: Evidence with adolescents from Switzerland

How to cite

Tyahlo, Svitlana; Maillard-Bjedov, Tjaša; Huber, Martin; Herz, Holger. 2017. "How language affects economic behavior: Evidence with adolescents from Switzerland." AEA RCT Registry. AEARCTR-0002021. https://doi.org/10.1257/rct.2021-4.0

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

Svitlana Tyahlo* , Tjaša Maillard-Bjedov , Martin Huber , Holger Herz

* Primary investigator

Registry ID
AEARCTR-0002021
DOI
https://doi.org/10.1257/rct.2021-4.0
First registered
Last updated
Status
Completed
Trial dates
2017-04-03 – 2017-04-05
Countries
Switzerland (Fribourg)
JEL codes
D03, D14, D90, Z13
Registry record
http://www.socialscienceregistry.org/trials/2021

Abstract

The linguistic-savings hypothesis (see Chen, M. K., 2013: “The Effect of Language on Economic Behavior: Evidence from Savings Rates, Health Behaviors, and Retirement Assets”, American Economic Review, 103(2), pp. 690-731) states that languages that grammatically associate the future and the present foster future-oriented behavior in terms of savings and other economic outcomes. Our study primarily investigates differences in time preferences between French and German speaking students in a secondary school in the Swiss canton of Fribourg. We conduct a pen-and-paper study with real payoffs to test (i) whether there are differences between French and German speakers in exponential discount rates (δ) and (ii) whether there are differences between French and German speakers in present bias (β), and (iii) whether the bilingual exhibit lower (larger) time discounting than French (German) speakers.

Intervention

We do not implement a classical intervention, since our hypotheses are based on natural language differences between subjects.

Experimental design

A pen-and-paper based study takes place during regular school hours in a bilingual secondary school in the Swiss canton of Fribourg. Pupils voluntary participate in the study involving incentivized choice experiments for time and risk preferences. In addition, we collect important sociodemographic characteristics (e.g., gender, year and months of birth, migration background, parents’ background, language exposures, family structure and material conditions, schooling, plans for the future, values) that are likely to correlate with the study language and outcomes of interest.

Design details

Randomization method
We do not implement a classical randomization, since our hypotheses are based on natural language differences between pupils.
Randomization unit
We do not implement a classical randomization, since our hypotheses are based on natural language differences between pupils.
Clusters
The number of clusters corresponds to the number of classes represented in the study. There are 20 German-speaking classes and 9 French-speaking classes yielding in total 29 clusters conditional on full attendance.
Observations
Total of 561 observations conditional on full attendance.
Arms
381 pupils in the German and 180 pupils in the French part of school conditional on full attendance.
Primary outcomes
We measure time preferences using 4-week exponential discounting rate δ and present bias β.

Cited by

No citation data available for this registration yet.

Keywords