This data contains the collected information of the survey experiment that was carried out within the DISCEFRN project (see metadata section Funding Information).

Within DISCEFRN, we combined web-scraped job vacancy data of the Norwegian labour market with a factorial survey experiment that exploits real-world variation in CEFR requirements within these ads (n vignette ratings= 10,495; n employers= 1,527) to examine whether fictitious applicants with a refugee background face less language-based discrimination on the individual level among employers who use standardized language requirements in their (real-world) ads compared to those that don’t. We thereby varied different applicant characteristics related to ethnic origin and to formal (CEFR certificate) and informal language indicators (e.g. spelling, argumentation, professional reference on unobservable relational skills) within vignettes and collected information on job-, firm- and employer characteristics (most notably attitudes towards different refugee groups) with standard survey items. This allowed us to assess whether CEFR requirements are primarily mitigating biased applicant evaluations that are related to language-based statistical/error discrimination (less relevance of informal language indicators), related to discrimination tastes (less relevance of group-related attitudes), or both.

This dataset contains all information on the survey experiment. It is a stand-alone dataset and contains all relevant data to re-produce associated publications (See metadata field on Publications) or be reused for other research interests. Yet, it can still be linked to additional DISCEFRN datasets, i.e. the web-scraped data set, that also holds information on those employers that did not participate in the survey experiment (https://doi.org/10.18710/K6WA0V).

STATA, 17