Effects of Multiple Exposure on Elicited Imitation Test Administration
Over the past 4 years we have examined the effectiveness and practicality of elicited imitation (EI), as a method of testing L2 proficiency. EI is based on the theory that a person’s ability to reproduce a sentence is linked with his understanding of that language's structure. During an EI test the subject hears a sentence, forms a cognitive representation, and then produces a sentence by accessing that cognitive representation (Bley-Vroman & Chaudron 1994). Our previous research showed that EI results predict Oral Proficiency Interview scores within 2 sub-levels (Lonsdale et al., 2009). In this paper we will examine the effect of multiple exposure on the results of our EI test.
We have administered 7 different test forms to 1279 subjects, 361 of which have taken it multiple times. We first categorized the subjects according to how many times they took the test: 2, 3, or 4 times. Before looking at the overall effect of multiple exposure, we first grouped these subjects by language learning level and L1, and then analyzed their scores across multiple tests to investigate other possible sources of variance. We used the placement tools of the English Language Center, including LAT Speaking, ECT Speak and SOPIs, to establish a baseline of competence to compare with subject's performance on EI administrations. By comparing subjects' performance for different degrees of exposure, we were able to determine the degree to which multiple exposure impacted the results of an EI test.
[Jeremiah McGhee, Malena Weitze and Dan Dewey]