July 22, 2024Journal of Educational MeasurementOpen Access

Using Automated Procedures to Score Educational Essays Written in Three Languages

View Full Paper

Ask AI

Bookmark

Automated essay scoring achieved higher agreement with human raters when utilizing LaBSE compared to multilingual BERT across German, Italian, and Czech essays.
Evaluations using the Common European Framework of Reference showed that LaBSE accurately predicted scores across the majority of proficiency levels.
Distinct sentence embedding architectures directly influence scoring consistency, supporting the deployment of language-agnostic models in educational testing.

Authors

TFTahereh FirooziUniversity of Alberta HMHamid MohammadiUniversity of Alberta MGMark J. GierlUniversity of Alberta

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Comparative evaluation reveals superior automated essay scoring accuracy with LaBSE over mBERT across three languages, highlighting the value of language-agnostic models in educational testing.

Cite This Study

Firoozi et al. (2024) studied this question.

synapsesocial.com/papers/68e5f73ab6db64358758bd5b https://doi.org/10.1111/jedm.12406

View Full Paper

Ask AI

Bookmark

Implication

Comparative evaluation reveals superior automated essay scoring accuracy with LaBSE over mBERT across three languages, highlighting the value of language-agnostic models in educational testing.

Using Automated Procedures to Score Educational Essays Written in Three Languages

Key Points

Discussion

Implication

Cite This Study

Using Automated Procedures to Score Educational Essays Written in Three Languages

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider