104 11 3259-279 2012-2013 2012 2013 2012 2013 2012 2013 2012 2013
260 1 4000 1
261 Computerized adaptive testing, CAT item response theory, IRT 2004 Wang & Chen, 2004Wang, Chen, & Cheng, 2004 2011 2012 2013 1992 IRT Lord, 1980 IRT Rasch Rasch, 1960 Wright & Mok, 2000 1 P ni exp( θn bi) = 1 1 + exp( θ b ) n i P ni n i θ n n b i i
262 1 1 1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.1-3 -2.6-2.2-1.8-1.4-1 -0.6-0.2 0.2 0.6 1 1.4 1.8 2.2 2.6 3 1 Rasch Rasch CAT IRT ConQuest Wu, Adams, & Wilson, 1998 Rasch IRT IRT ConQuest
263 Conquest IRT -3~3-4 3.5 Rasch 3 3 Rasch -3-3 ConQuest unweighted mean square errorweighted mean square error 1 0 1.0 Smith, 2000 0.7~1.3 misfit itemswright & Linacre, 1996 1.0 Rasch
264 1. 2. 3. Birnbaum, 1968 600 600 Rasch 0 1.5 1. Rasch 2. 3. 3 Expected a Posterior, EAPEAP 2 I u ni 1 u ni n1 n2 nl θn = i i θn i θn Lu (, u,, u ) P( ) (1 P( )) 2 P ni θ n 1u ni n i u ni 1 0 L u θ n 3 ˆ = = θ EAP 61 [ Q ( ) ( )] 1 r L Qr W Q r r 61 LQ ( r) W( Qr) r = 1 3 θ EAP Q ˆEAP r r W Q r
265 r 31 1 Q 1-3 W Q 1 0.00443EAP maximum fisher informationirt information 4 2 P i Ii ( θ ) = 4 PQ i i P i P i θ 4 standard error, SE 5 1 SE( θ) = 5 I( θ ) 45 0.5 0.5 θ θ 4
266 2 20 0.5 20 0 1 0.3 0.9 1.5 0.5 0.9 300-1000 0.9 1. 2. Rasch 3. 4. 1 0 Rasch -1.08 0.25
267 2.46Rasch 0.25 2 0.6 1.63 1 1100205 IRT -0.15 51 θ -0.150.482615352625979 θ b SE C095010650 1 A A 1 0-1.081 0 09:22:17 09:22:40 C101070220 2 C C 1 0.25 0.25 2.45947 09:22:40 09:23:09 C101100150 3 A C 0 0.6 0.604 1.63337 09:23:09 09:24:09 C101110140 4 B B 1 0.3 0.3 1.23592 09:24:09 09:24:43 C099010150 5 A C 0 0.55 0.549 1.06763 09:24:43 09:26:00 C099020090 6 D B 0 0.3 0.301 0.93216 09:26:00 09:27:06 C101050340 7 D D 1 0.1 0.1 0.84639 09:27:06 09:27:45 IRT -4~4 IRT -4~3.5 logit 0~100 3.5 100 3.425 99 3.3 3.275 97 2 2-4~3.5 3.5 100 3.425 99 0.2 56 0.1 55 0 53-0.1 52-0.2 51-3.85 2-3.925 1-4 0
268-4 3.5 2 2 2012 10 2013 10 701 3 3 2012 2013 93 97 82 91 89 82 82 85 346 355
269 3 1. 5. 2. 4. 3. 6. 3 4 2012 2013 701 17-18 20 0.5 9-15 11 2012 2013
270 4 2012-2013 4 2012 4 2012 2012 θ θ θ 93 1.35-1.3 0.117 0.543 82 1.2-1.85 0.125 0.479 89 1.35-1.35 0.108 0.551 82 1.3-1 0.173 0.480
271 4 2012 0.1 0.17 55 0.48-0.55 95% 1.28-1.0 0 3.4-4 3-3 2012 2012 one way anova 5 5 2012 SS MS F P- 0.210678 3 0.070226 0.260666 0.853709 2.63102 92.13817 342 0.26941 92.34885 345 5 p > 0.05 2012 2013 2013 6 2013 2013 θ θ θ 97 1.45-1.0 0.156 0.453 91 1.2-1.05 0.134 0.462 82 1.4-1.0 0.257 0.485 85 1.1-1.15 0.167 0.429 6 2013 0.13 0.26 55-57 0.43-0.48 95% 1.21-0.85 0 3.4-4 1 2-2 2013 2012 2013 one way anova 7
272 7 2013 SS MS F P- 0.758887 3 0.252962 1.082237 0.356596 2.630347 82.04284 351 0.23374 82.80173 354 7 2013 p > 0.05 2012-2013 2012 2013 99-101 8 99 101 t-test p-value 8 2012-2013 2012 101 100 99 2013 2012 2013 2012 2013 0.119 0.134 0.134 0.170 0.069 0.165 87 65 77 t-test p 0.808 0.562 0.149 * 8 2012 2013 p > 0.05 9 3 3
273 0.8-0.8 9 2012-2013 99 99 3 4 / 0.8,-0.8 p 0 7 0 1.582 0.453 0 60 3 0 6 1 p p > 0.05 0.5 0.5 2013 355 6559 100 19 10 10 2013 100 1 C101070220 258 161 0.624 2 C101110140 214 42 0.196 3 C101010500 159 78 0.491 4 C101010510 159 103 0.648 5 C101010520 159 136 0.855 6 C101120190 142 101 0.711 7 C101050030 138 81 0.587 8 C101040020 134 96 0.716
274 9 C101050050 126 54 0.429 10 C101050340 126 54 0.429 11 C101100150 120 55 0.458 12 C101040040 119 78 0.655 13 C101070070 119 78 0.655 14 C101100320 117 52 0.444 15 C101010130 109 70 0.642 16 C099020170 102 54 0.529 17 C101030280 102 35 0.343 18 C101050190 102 22 0.216 19 C099020200 101 64 0.634 19 80% 20% 0.5 0.12 0.78 0.25 0.05 10 1 5 18 1 C101110140 A B C D 5 C101010520
275 A4 1 2010 A B C D 18C101050190 A B C D
276 2004 2012 2013 2012 2013 2012 2013 2012 2013
277 2004Rasch 27 4 637-694 19929 15-9 2004 Birnbaum, A. (1968). Some latent trait models and their use in inferring an examinee s ability. In F. M. Lord & M. R. Novick (Eds.), Statistical theories of mental test scores (pp. 397-479). Reading, MA: Addison-Wesley. Lord, F. M. (1980). Applications of item response theory to practical testing problems. Hillsdale, NJ: Lawrence Erlbaum Associates. Rasch, G. (1960). Probabilistic models for some intelligence and attainment tests. Copenhagen, Denmark: Danmarks Paedogogische Institut. Smith, R. M. (2000). Fit analysis in latent trait measurement models. Journal of Applied Measurement, 1(2), 199-218. Wang, W. C., & Chen, P. H. (2004). Implementation and Measurement Efficiency of Multidimensional Computerized Adaptive Testing. Applied Psychological Measurement, 28(5), 295-316. Wang, W. C., Chen, P. H., & Cheng, Y. Y. (2004). Improving Measurement Precision of Test Batteries Using Multidimensional Item Response Models. Psychological Methods, 9(1), 116-136. Wright, B. D., & Linacre, J. M. (1994). Reasonable mean-square fit values. Rasch Measurement Transactions, 8(3), 370. Wright, B. D., & Mok, M. (2000). Rasch models overview. Journal of Applied Measurement, 1(1), 83-106. Wu, M. L., Adams, R. J., & Wilson, M. R. (1998). ACER ConQuest: Generalized item response modelling software. Melbourne, Australia: Australian Council for Educational Research. Chi-Yi Hsieh is an Associate Professor of the Department of Chinese Studies, National Quemoy University, Kinmen County, Taiwan. (Corresponding Author) 103 05 27 103 11 30 103 12 01
National Taiwan University of Science and Technology Journal of Liberal Arts and Social Sciences 2015, 11(3), 259-279 READING COMPETENCY TESTING AND RESULT ANALYSIS FOR STUDENTS OF APPLIED CHINESE MAJORS- A CASE STUDY ON THE STUDENTS OF THE DEPARTMENT OF APPLIED CHINESE OF WENZAO URSULINE UNIVERSITY OF LANGUAGES Chi-Yi Hsieh Department of Chinese Studies, National Quemoy University ABSTRACT This research aims to explore the reading competency of undergraduate students majoring in applied Chinese related specializations by utilizing the existing computerized Chinese competency testing system of Wenzao Ursuline University of Languages (Wenzao University). The experiment was implemented on the students studying at the Department of Applied Chinese of Wenzao University in 2012 and 2013 (tested students), targeting their reading abilities. For testing tools, the computerized Chinese competency testing system set up at Wenzao University was deemed more effective and reliable than the general pen-and-paper testing methods. Results of the reading competency tests implemented in 2012 and 2013 showed that the tested students fell rather concentrated in the middle section of the scale. In addition, when viewed from collective performance of each grade, there were slight differences between the four grades, but none of the differences reached the level of statistical significance. This shows that there were no differences between the four grades in terms of reading abilities. The same batch of students was tracked from 2012 to 2013 and each student was analyzed for his/her reading abilities at different time sectors (advancing from one grade to the next). We found that no significant differences were present in reading abilities when a student moves from one grade to the next. Further analysis was conducted on three groups of mixed grades labeled high, general and low reading abilities, and no significant differences were found in between the results of 2012 and 2013.
Reading Competency Testing and Result Analysis for Students of Applied Chinese Majors 279 Finally, this research conducted an analysis on the differences between students of the Department of Applied Chinese and students of other majors. Significant differences were only found in three sections: one ranked in the scale as easy and two difficult. The former involves knowledge of procedures on Chinese writing and layout, and the latter involves meanings of vocabularies and meta rules of classical Chinese texts. Keywords: Chinese reading abilities, Department of Applied Chinese of Wenzao Ursuline University of Languages, computerized Chinese Competency Test