Essay tests are widely used to assess ESL/EFL learners' writing abilities for instructional, administrative, and research purposes. Relevant literature was searched to identify 70 empirical studies on ESL/EFL essay tests. The majority of these studies examined task, essay, and rater effects on essay rating and scores. Less attention has been given to the effects of examinee factors, scoring methods, and assessment contexts. This absence seems mainly to be the result of a traditional concern with controlling for task and rater variability as "sources of measurement error." This article argues for viewing these factors as "sources of variability" that contribute to the richness and uniqueness of the contexts within which writing performance and assessment occur and for taking them into account when interpreting and using essay test scores. The paper concludes with several implications for research and practice.