Abstract

The has become an increasingly popular mode of testing, a staple of more and more placement and proficiency tests. I The following scenario is common: A student receives a topic and the specifications for a piece of writing to be completed in twenty or thirty or forty minutes. The student writes furiously (I mean this in at least two senses). The completed essay is bundled with dozens or hundreds of others and delivered to a roomful of teachers, all cranky with anticipation of their task but glad to pick up some extra money for their pains, who spend about a minute on each essay, reducing it to a 1, 2, 3, 4, 5, or 6 while they watch the clock and wait for the lunch break, and occasionally chuckle together over a particularly dumb and funny student sentence. Sometime later, the student receives notification of whether she is competent, or proficient, or literate, or whatever label this enterprise is supposed to impose. The absurdity of all this is apparent.2 Good writing comes from a writer with something important to say to an interested reader. But committed writers and interested readers are nowhere in this scenario. It would appear that testing writing means asking students who would rather not write to produce something for readers who would rather not read. Without commitment and interest, all concerned must settle for efficiency: writers try to finish off a topic in minutes, and readers try to finish off each essay in seconds. While complaining about the perversion of writing and reading into a disinterested rush, I must also admit that producing an acceptable writing sample does require, however scarcely, those skills exploited more fully in most academic writing tasks. The writer being tested must invent content, focus and form it into sentences and paragraphs (keeping audience and purpose in mind), revise, and edit. Because writing samples mimic the production of an academic essay, we can conclude that the writing sample test, with all its shortcomings, does test skills relevant to academic writing.3 This conclusion introduces the central problem I wish to consider. When tests equate proficiency, competency, and literacy with writing academic essays, they maintain a severely understated and mistaken