The impact of information and role models on garment workers in Vietnam: Evidence from a field experiment

How to cite

Yao, Xinjue; Hensel, Lukas; Chen, Yuyu. 2023. "The impact of information and role models on garment workers in Vietnam: Evidence from a field experiment." AEA RCT Registry. AEARCTR-0012457. https://doi.org/10.1257/rct.12457-1.1

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

Xinjue Yao* , Lukas Hensel , Yuyu Chen

* Primary investigator

Registry ID
AEARCTR-0012457
DOI
https://doi.org/10.1257/rct.12457-1.1
First registered
Last updated
Status
In development
Trial dates
2023-11-21 – 2024-04-10
Countries
Viet Nam ()
Registry record
http://www.socialscienceregistry.org/trials/12457

Abstract

Managers at a garment factory in Vietnam complain about low rates of productivity and high rates of turnover among their workers. Based on qualitative research, we found that workers invest little in acquiring new skills despite there being clear monetary payoffs. We hypothesize that this has two reasons: First, workers are not aware about the returns to investing in new skills. Second, they see existing skilled and productive workers as being very different from themselves and don’t perceive high skills to be possible and desirable for them. We plan to experimentally test whether relaxing either constraint separately or relaxing them jointly affects workers’ investment in their skills, their effort, their productivity, as well as their retention and absenteeism rates.

Intervention

The intervention is designed to mimic the typical social media content that workers consume in their spare time. Each intervention consists of eight short videos that will be delivered electronically in the Facebook groups set up at the beginning of the study. Workers will be allocated to one of four groups: information intervention, role model intervention, combined intervention and placebo group.

Experimental design

We will test two interventions aimed at increasing workers’ skill acquisition, increasing their effort, and, ultimately, their productivity. The intervention is designed to mimic the typical social media content that workers consume in their spare time. Each intervention consists of eight short videos that will be delivered electronically in the Facebook groups set up at the beginning of the study. We will incentivize workers to watch the videos by sending out weekly follow-up surveys with prizes for respondents who answer questions about the content of the videos correctly.

Design details

Randomization method
Randomization happens in two stages. First, we randomize 59 production lines into high- and low-intensity lines. 2/3 of the production lines will be high intensity lines while 1/3 will be low-intensity lines. On high-intensity lines 90% of workers receive any of the three treatments. On low-intensity lines only 50% of workers receive any of the three treatments. This is calibrated to allocate 25% of workers to the placebo group overall. Second, within each line we randomize which workers are in which group. All three treatment groups will have the same size on each treatment line. This ensures that each treatment group will contain 25% of sampled workers. Randomization will be implemented using rerandomization until balance on baseline primary outcomes and demographic characteristics is achieved.
Randomization unit
Individual
Clusters
The treatment is not clustered.
Observations
2200 individuals
Arms
550 workers control, 550 workers information intervention, 550 workers role model intervention and 550 workers combined intervention.
Primary outcomes
Investment and usage of skills, effort measures, productivity
Secondary outcomes
See attached file.

Cited by

No citation data available for this registration yet.

Keywords