The survey reported here is part of a larger project targetting placement practices in language programs. The overall goal of the project is to develop a manual of best practices in placement testing that would be of use to those people involved in the placement process. Although in one sense good testing practices transcend specific testing purposes and any training in test development would be useful for placement testing, we wanted to tailor the manual as much as possible to concerns and issues common in assessing language for placement decisions. In order to begin the tailoring process, it was necessary to first get some information on what programs are currently doing for placement. A survey seemed like the most efficient way to accomplish this initial goal, and this paper reports on the development and results of that survey.
We recognized from the outset that we did not have the resources nor the need to produce a survey of placement practices that would allow us to state with confidence the precise percentage of programs using one procedure versus another or state definitively that statistically significantly more programs assessed reading than speaking. The types of test development procedures expounded in the testing literature would be useful across a range of placement procedures and good testing practices are at the heart of good placement testing practices. What we did need, however, was a kind of "slice of life" view of what was happening in a fair number of programs so that we could develop a manual that would discuss those good testing practices within the context of a recognizable (to the reader) placement process.
With this in goal mind, five research questions were posed:
It was hoped that answers to these general questions would give a good stating point for thinking about how to develop the placement testing manual. If programs tended to focus on productive skills such as writing and speaking, for example, then a thorough discussion of the use of raters and rating scales for assessment would be necessary. On the other hand, if most programs were using selected-response tests, then focusing only on performance type assessments would be doing them disservice. By the same token, it would make no sense to pitch the handbook to language testing professionals if the people responsible for placement testing in most programs are teachers. This is not to say that the manual should support the status quo no matter what that is, but rather that for the handbook to be useful, it must address concerns of real programs while recognizing their constraints. Discovering the placement concerns of a number of real programs would obviously be very helpful in planning the manual.