The present study reports the performance of 63 senior non-native undergraduate university students on C-Tests, a decontextualised C-Test, a spelling test, a matching vocabulary test (MVT) and a disclosed Test of English as a Foreign Language (TOEFL). It was postulated that the availability of the first half of deleted words might render the items independent from the context in which they appear. It was also postulated that the directions given on the C-Tests might call for the test takers’ spelling knowledge rather than their vocabulary knowledge and language proficiency. To test the hypotheses, the C-Tests designed by Klein-Braley (1997) were administered along with their decontexualised versions. The analysis of the data showed that the C-Tests correlated significantly with all the tests except the decontextualised C-Test, implying that context plays an indispensable role in answering C-Tests. Although the C-Tests correlated significantly with the MVT and TOEFL, their coefficients were not high enough to establish them as independent measures of vocabulary and language proficiency. The initial principal component analysis revealed three factors. While the C-Tests had the highest loadings both on the first and second factors, the decontextualised C-Test and the spelling test loaded on the third. Varimax rotation showed that the MVT and TOEFL had the highest loadings on the first or G-factor whereas the C-Tests loaded on the second. The decontextualised C-tests had the highest loading on the third factor. These findings show that performing on the C-Tests requires an ability other than spelling ability, vocabulary knowledge and language proficiency. This ability is method specific and depends on understanding the context in which C-Test items appear.

Keywords