EIKEN(STEP) research

Also today, i browsed several articles featured in STEP bulletin.

What is the most touchy and interesting article is by Dr. Akiyama, about evaluation of speaking test using Item Response Theory(IRT). What he read to research this topic, contains what i have read in Language testing class, and on my own. Moreover, his topic-evaluation of test he conducted-is just what i chose as s topic as my class paper! I would like to read some articles he read and i have not yet read. Also, his research encouraged me to research about this topic regardless of using IRT or CTT(classical testing theory).

Several articles i read are about vocabulary and test scores, and use correlation or ANOVA. Corpus analysis is also used to research vocabulary, but it would be hard to make a picture many audience understand. Correlation is used for comparing vocabulary test scores and proficiency tests scores, and Anova is used for comparing vocabulary learning strategy and proficiency tests scores.

For me, it seems that sophisticated stats analysis such as Manova, Factor Analysis, Structural equation model, covariance structural analysis, and IRT can make audience understand, because it seems to draw a big picture. For example, Dr. Okada's aritcle,Dr.Matsumoto's article and Ms.Suzuki and Dr.Kumazawa's article seem to succeed in showing a big picture for audience to understand them.