|
Classical test theory exists limitation in theory system, it takes true score model as its theory frame(Observed Score equals True Ability plus Random Error, the math equation is X = T + E). various criterions it adopts, such as difficulty, discrimination and reliability , are sample dependent . That is to say, these criterions will change as test sample, therefore, the same test paper hardly acquires accordant difficulty, discrimination and reliability. Item response theory is the most advanced test theory at present, its naissance based on Classical test theory’ s limitation. Its item parameters such as difficulty, discrimination and reliability, are sample-free, That is to say, these parameters will not change as test sample changes. S-P chart theory is a branch of item response theory, an ideal S-P chart should all be 1 on the left of S line and should all be 0 on the right of S line, on the other hand, the ideal S-P chart should all be 1 on the top of P line, and should all be 0 underneath P line. If not this, it appears that there are some problems in teaching, these problems can be represented by alarm coefficient. The teacher can quickly find out his students’ study difficulty and learning achievement, or else, he can quickly find out the question proper or not, in addition, there are some problems in his teaching. Our quality analysis about Standardization examination mainly bases on pass rate, average score, excellence rate, and so on, a great deal of useful information is lost in students’ papers. By computer programming s-p chart, we can analyze the whole paper, and feedback the situation of studying and teaching in time.
|