| The goal of the present study was to investigate the validity issue in vocabulary testing that had been repeatedly raised for more than half a century, and further explore the possibility of approaching a theory of the ESL lexicon to serve as a defensible construct for vocabulary testing and a useful guideline for vocabulary teaching in TESOL.;The present study began by examining the validity issue raised by two studies, one by Sims in 1929 and the other by Henning in 1991. Both Sims and Henning were concerned with the quality of vocabulary tests of their times, but they failed to define what "the same thing" was that the tests they were evaluating were testing before resorting to the method of inter-correlations to evaluate them. While Sims concluded that the checking test was not a valid one because it did not correlate well with other tests being evaluated, Henning believed that the current TOEFL vocabulary test was a valid one because it correlated well with other multiple choice variations designed by him and his associates.;The analysis of more than 800 vocabulary test items collected from most representative ESL/EFL tests throughout the world in light of the relational model of the lexicon proposed in recent years in cognitive science revealed that the synonymous relation was overwhelmingly privileged in test construction. The study found no support of the claims made by either Sims or Henning. It did, however, seem to support the idea of testing vocabulary in reading contexts which would have the potential to include a variety of lexicosemantic relations and engage the examinees in a more productive way to handle the task at hand.;The present study also discussed the possible steps that could be taken in order to make the borrowed theory work in TESOL. |
