Abstract
This study was carried out to investigate raters’ relability and consistency when assessing writing performance of EFL learners of Vietnamese at a southwestern university of Vietnam. Data were derived from a group of four raters who were asked to rate 50 essays written by the same number of third-semester university students. The researcher applied Rasch measurements to analyse the data using the FACETS software program. The results from the analyses of three and four raters showed that a considerable proportion of candidates was misfitting, suggesting problems with the scale used to rate essays and its implementation by raters. Also, most raters were not consistent (one was inconsistent and two tended to be overconsistent) in their ratings; the same tendency was found amongst criteria. The overall results suggest a call for rater-training through which rater consistency is improved, and also a call for adjustments on the scale by redesigning and, if necessary, eliminating poorly-constructed criteria, or replacing with a better-functioning rating scale.
Recommended Citation
Phung, De Van
(2013)
"Rater Behaviors in Assessing EFL Compositions,"
CNU Journal of Higher Education: Vol. 7:
Iss.
1, Article 9.
DOI: https://doi.org/10.70997/2546-1796.1091
Available at:
https://jhe.researchcommons.org/journal/vol7/iss1/9