英语论文网

Chapter five Results, Discussionand Implications

5.1 Results and Discussion
In this study, the author wanted to explore whether SAQs were more valid thanMCQs in testing reading comprehension, that is, we wanted to find out if SAQs weremore valid than MCQs to be employed in measuring reading ability, and results wereanalyzed in order to find whether our hypothesis was a practical one. We clarified theproblem by using multiple methods of analyses. The following approach was adopted toensure the reliability of the findings.First, through two-way ANOVA, we found out, test approach’s main effect was notsignificant, that is, there was no significant difference between the two test approaches,MCQs and SAQs, in participants’ reading comprehension performance; second, the maineffect of participant’s English proficiency level on testing reading comprehension wassignificant, that is, there were significant differences among three subgroups of eachrandom group in their reading comprehension performance. And third, there was aperceived significant interaction between test approach and proficiency level, that is,there was an interactive effect of proficiency level in relation to participants' reading testperformance perceived in the two test approaches, MCQs and SAQs. From these, wecould conclude that, within two test approaches, MCQs could not show significantdifference in distinguishing good, fair and poor students while SAQs could showsignificant differences among good, fair and poor students.
…………

Conclusion

Language testing is indispensable to language teaching because it is one of the mostimportant measures for checking the outcome of language teaching (Bachman 1997，Liu2000，Li 2001, etc.). The validity and reliability, however, are always the concern oflanguage testing researchers because it is not easy to ensure high validity and reliabilityat the same time. Therefore, language-testing researchers should try their best tominimize the measurement error. In this study, some fruitful conclusions can be drawnfor language testing and teaching.Theoretically, on the one hand, the short-answer type is the most valid in terms offace validity and response validity; on the other hand, the completion type is the mostvalid in terms of concurrent validity and predictive validity. The MC type, however, isproved to be the least valid in any case. Content validity and construct validity beingequal, face validity and response validity, most related to item types, and another twotypes of validity, concurrent validity and predictive validity involved in the validationprocedure of Alderson et al (1995), become the most important indices for reading tests.The comparatively higher validity means that the corresponding type of reading test ismore suitable for testing reading ability in general.
……………
Reference (omitted)