Font Size: a A A

A study of test-type related variability of interlanguage performance among Japanese EFL learners: A focus on relative clause tests

Posted on:2000-03-19Degree:Ph.DType:Dissertation
University:Hiroshima University (Japan)Candidate:Ito, AkihiroFull Text:PDF
GTID:1465390014462401Subject:Education
Abstract/Summary:
The purpose of this study is to investigate (1) the effects of differences of relative clause test-types on the accuracy rates in interlanguage performance, (2) the reliability and validity of four different test-types and how many items each test-type should contain in order to be very reliable and (3) how the learning order of relative clause type changes in line with the difference of data elicitation techniques.;Four 24 item relative clause tests with four different test-types (Translation, Cloze Procedure, Grammaticality Judgment, and Sentence Combining) were administered to 120 Japanese learners of English.;The conclusions drawn from this study are (1) the combination of Tarone's interlanguage continuum model and DeKeyser's proposal can be regarded as a valid predictor of the test-type related interlanguage variability of Japanese EFL learners; (2) the magnitude of variability has a tendency to become smaller as the proficiency level increases; (3) of the four relative clause tests, Sentence Combining test-type, has the highest potential to be a reliable and valid measurement of relative clauses; (4) 12 to 82 items should be added afresh to the four relative clause tests for them to be made very reliable; (5) the learning and/or acquisition order of relative clauses changed according to the function of test-types used as the elicitation device; (6) the relative clause type genitives (GEN) changed its position in difficulty order among the selected relative clause types in Grammaticality Judgment; (7) because of the difference of test-types, the difficulty or learning order of English relative clauses varies as follows: Translation: SU>DO>GEN>OPREP: Cloze Procedure: SU>DO>OPREP>GEN: Grammaticality Judgment: SU>GEN>DO>OPREP; and Sentence Combining: SU>DO>OPREP>GEN; and (8) the lower reliability and validity of Grammaticality Judgment was explained in terms of the problems with the process of test construction, and the author has proposed the idea that the function of 'whose + NP' unit should be taken into consideration in order to increase the internal consistency in GEN items in relative clause tests, especially in Grammaticality Judgment test.
Keywords/Search Tags:Relative clause, GEN, Test-type, Grammaticality judgment, Interlanguage, Order, Learners, Japanese
Related items