Subtitling vs. Dubbing: How Out-of-School Learning Shapes English Proficiency
Baumeister, Hanushek & Wößmann: "Out-of-School Learning: Subtitling vs. Dubbing and the Acquisition of Foreign-Language Skills" CRC Discussion Paper No. 536
The development of English-language skills, a near necessity in the global economy, is commonly attributed to classroom instruction. However, new research by Frauke Baumeister (ifo Institute), Eric A. Hanushek (Hoover Institution, Stanford University), and Ludger Wößmann (LMU Munich, Project A06) demonstrates the overwhelming influence of out-of-school learning factors such as historical national decisions regarding TV content translation. Their analysis focuses on whether non-English speaking countries in Europe choose subtitling, which maintains the English audio track, or dubbing into the local language, which removes it. To isolate the causal effect of translation mode, the authors employ a cross-country between-subject model. This approach compares English skills to math skills, using math as a counterfactual assumed not to be affected by TV translation mode. This methodology allows the researchers to control for populations’ overall skill levels that might otherwise bias simple comparisons.
The central finding is a large positive effect of subtitling on English-language skills. Across various proficiency measures—including the English Proficiency Index (EPI), the Test of English as a Foreign Language (TOEFL), and the representative Adult Education Survey (AES)—the estimated effect generally exceeds one standard deviation (SD). The baseline estimate for EPI scores is 1.4 SD, which corresponds to the average difference in English proficiency between dubbing countries like Germany and subtitling Scandinavian countries. Estimates for the representative AES are similar in size, indicating that results are not driven by selective test-taking. Furthermore, effects are strongest for the TOEFL listening and speaking skill domains, while the effect on TOEFL reading skills is statistically insignificant, in line with oral learning from subtitled TV. These results are robust to accounting for potential confounding factors like similarity between local languages and English, size of country populations and language communities, or the starting age and instruction time of foreign-language teaching in schools.
The estimates are best interpreted as long-run effects of employing subtitling for multiple generations rather than the effect of individual TV viewing. They underscore that national media policies have powerful impacts on a population’s skill accumulation, suggesting that educational policies must consider broader, non-school ways of learning.
Link (pdf): Out-of-School Learning: Subtitling vs. Dubbing and the Acquisition of Foreign-Language Skills
