This study investigated the effect of writing task topic on learner performance in a second-language writing test, in this case the Michigan English Language Assessment Battery designed to test proficiency in English as a Second Language. The 64 topics or "prompts" used in the test (offered as pairs of options) were categorized according to writing task type (expository/private; expository/public; argumentative/private; argumentative/public; and a combination of two or more of the previous types) and the categories assigned a level of difficulty. Scores received on the test were then correlated with topic difficulty. Contrary to expectation, the mean writing score increased rather than decreased as topic difficulty increased. Implications for test construction and for rater judgment are examined. (MSE)