Verb argument construction complexity indices and L2 writing quality: Effects of writing tasks and prompts

The purpose of the study is to investigate whether verb argument construction (VAC) based indices of syntactic complexity significantly vary in different second language (L2) writing tasks and whether such complexity measures vary in their relationship to writing quality. The present study used data from three different L2 writing tasks: descriptive, independent, and integrated. Descriptive essays (N = 70) were taken from the Michigan State University corpus (Connor-Linton & Polio, 2014). Independent (N = 70) and integrated (N = 70) essays were selected from TOEFL public use dataset. Each corpus contained an equal number of essays written on two prompts (for a total of six prompts). Each essay was rated by two trained raters using TOEFL independent or integrated writing rubrics. The selected essays were stratified by writing scores, and it was made sure that the scores were normally distributed. The Tool for the Automatic Analysis of Syntactic Sophistication and Complexity (Kyle, 2016) was used to calculate the VAC-based indices. Multinomial logistic regression and linear models were used to analyze the data. The findings indicate that VAC-based complexity measures vary by L2 writing tasks and that the relationship between VAC measures and L2 writing quality is also task-dependent with few prompt-based effects.