This study examines changes in the scores and linguistic characteristics of learners' responses to TOEFL-iBT writing tasks before and after English-language instruction to address two validity questions: whether changes in learners' overall English language proficiency are associated with changes in the linguistic characteristics of their essays and whether changes in essay characteristics are reflected in changes in essay scores. Data will consist of about 960 essays written by a sample of 240 students in response to TOEFL-iBT independent and integrated writing tasks before and after 15-16 weeks of English-language instruction. A combination of computer analysis and manual coding will be used to analyse the essays in terms of five dimensions of L2 writing ability: grammatical (fluency, accuracy, complexity, lexis), discourse (coherence, cohesion, text structure), sociolinguistic (register), strategic (interpersonal metadiscourse), and content and source use (e.g., verbatim source use). The results will then be compared across tasks, proficiency levels, learner groups, and test occasions in order to examine how essay characteristics vary across proficiency levels at each test occasion and across occasions; whether and how task-, course- and learner-factors relate to differences and changes in essay characteristics within and across occasions; and how essay characteristics relate to essay scores within and across test occasions. The study can contribute significantly to explaining the meaning of TOEFL-iBT writing scores; examine whether essay scores are sensitive to differences and changes in learners' writing abilities; describe the language and discourse aspects and abilities engaged by TOEFL-iBT writing tasks; and provide important insights concerning the nature of L2 writing proficiency as it develops over time and how and why patterns of L2 writing development vary across individuals and contexts.