Validation of a German language placement test based on a modified C-test procedure

The C-test procedure, a variant of the cloze procedure, has long been advocated as an easy and effective method of assessing foreign language proficiency in a reliable fashion (Eckes & Grotjahn, 2006). For this reason, the Department of Germanic Languages and Literatures at the University of Illinois opted to employ this procedure for the development of a new placement test. This dissertation described both the development process and the challenges encountered during the validation of the test. Extant research on the C-test procedure was critically reviewed and problems with earlier research that are incompatible with current validity theory identified. Based on this, a modified procedure was proposed for the new placement test. This modified C-test procedure combined the specification-based test development process introduced by Davidson & Lynch (2002) with consensus-based content validation. For the modified procedure, content experts made deliberate changes to C-test source texts in an attempt to enhance criterion-referencing and collect content-related validity evidence.;Results based on four pilot tests administered over the span of two academic semesters suggested that: (a) the modified C-test procedure delivered enhanced linkage to the department's curriculum, and (b) psychometric qualities such as reliability of test scores were not materially affected by the modifications. The study concluded that a modified C-test procedure might be an effective language testing procedure that is more compatible with current validity theory than the original C-test procedure.