Mart Christine Johnston (Takushoku University, Japan)
Abstract
In my presentation, I will share the results of the readability assessment of a high-stakes proficiency test for Japanese high school graduates. My aim is to evaluate text analysis tool agreement and identify factors influencing text difficulty. Longer texts presented greater challenges, while emails were easier due to their brevity. I observed stronger correlations between average sentence lengths and text difficulty indicators than word counts. I also noted similarities between Flesch-Kincaid grade level and lexile reading levels, likely due to their reliance on sentence length. Overall, the texts seemed appropriate for students in grades 8-9 in the U.S. education system. While CEFR (assigned by Text Inspector) and CEFR-J (assigned by CVLA) levels showed some comparability, their correlation scores were moderate. Additionally, the New General Service List (NGSL) displayed weak correlations, focusing mainly on word frequency. These findings highlight the need for a comprehensive approach to accurately assess text difficulty and tailor materials for the intended audience.
Research (Video, 25 minutes)
Reading
General Interest
About the Presenter
Mart Christine Johnston has been an English teacher in Japan for various age groups for 10 years. Before coming to Japan, she also worked in Singapore for six years. Currently pursuing a PhD in language education, her dissertation centers on integrating commonly used phrasal verbs into a corpus. Her interests lie in language testing and methodologies.