Two of the fastest growing, and most technical, areas of applied linguistics are second language acquisition (SLA) and language testing. Perhaps because of the degree of specialization required of those doing research in either one, together with the advent of separate conferences and journals for each, dialogue between practitioners in the two fields has been sporadic at best, and nonexistent in many cases.

This state of affairs is clearly unproductive. Advances in SLA in recent years, for example, have thrown new light on such crucial notions for test developers as “acquisition” and “proficiency,” on learnability and teachability, and on the relevant criteria for student placement for the purpose of instruction, leading in some quarters to requests for tests that are “interlanguage-sensitive,” or “developmental,” and that take the individual learner as the unit of analysis. Yet many test designers remain unaware of such developments and continue to produce ever more sophisticated proficiency measures, of high reliability, but obtained via normreferenced, group-level statistics, and of questionable validity when used to cluster learners for teaching purposes. Advances in language testing, conversely, are sometimes a closed book for many SLA researchers, whose training too often lacks a sufficiently rigorous grounding in research methods in general, and in measurement in particular. The result can be studies that are theoretically innovative and insightful, but of questionable validity due to inadequate measurement of the variables of interest, including second language abilities and development.