计算机自适应测验软件模型研究
摘要
计算机化自适应测验是现代测验研究中的一个新领域。它对于提高测验
效率与质量有着重要意义。传统的测验，不论考生水平高低全部要接受同一
批题目的测验，结果对水平低的考生来说，那些难度大的题目，根本无法作
答，反而引起焦虑。对高水平考生来说，那些难度小的题目，又测不出考生
的真实水平，徒然浪费精力。
计算机化自适应考试（Computer Adaptive Test，简称 CAT）解决了主动适
应考生状况的“因人施测”问题。CAT的出现不仅打破了两千多年沿袭下来的
以纸和笔作为作答工具的方式而改为计算机显示屏呈现、键盘与鼠标进行作
答的方式，而且与传统的测验相比，测验思想也发生了巨大的变革：它通过
给每一个被试建立一个个人化的测验来达到更为准确的测量，因为项目的选
择是根据被试的能力水平定身度量而成的，因而被试所做的每一个题目的难
度都是与其能力相匹配的。也就是说，水平高的被试能够避免做到相当简单
的题目，而能力低的被试能够避免做到超出其能力范围之外的题目。从其测
验思路中我们不难看出CAT考试的众多优点：（1）测验效率高；（2）可比
性强。由于特定的实施方式与记分方式，CAT最终的分数可转化为可相互比较
以及解释的量尺。
目前，在一些发达的国家如美国，CAT已在教育测验、职业测量、人事
测评等领域中大显身手，如美国研究生入学考试（Graduate Record
Examination）、工商管理类研究生入学考试（Graduate for Management and
Administration Test）以及全美护士国家委员会资格考试（Nurse National
Comittee License Test）等都已采取了CAT的方式。不难看出CAT代表了
今后教育、心理测验发展的方向与重点。
但是，在我国，有关自适应测验的计算机软件极为少见，国家大学英语
四、六级考试委员会正着手这方面的研究，但目前还没有成品的软件模型问
OS差霎自SI 砒十堂o为丫
Y’。。。·。八！u工ER S 工HESIS
世，因此，深入计算机自适应测验的研究对于我国测验技术的发展具有重要
的意义。
本文第一部分从分析传统测验的问题出发，提出了计算机自适应测验的
意义和必然性，同时，对国内外计算机自适应测验发展的状况进行了说明。。
第二部分对计算机自适应测验的理论基础进行了分析，着重讨论了项目
反应理论以及本理论在测验工作中的作用。详细地分析了难度、区分度、信
度和效度等重要的概念。
第三部分提出了计算机自适应测验软件模型的设计方法和实施过程，同
时还列出了自适应测验中的一些重要的算法，包括施测算法、组卷算法和抽
题策略等。
第四部分分析了所完成的计算机自适应测验软件模型方案设计。对其中
的功能设计，测验过程及其实现等作了详细的叙述。
第五部分对本软件的测试结果作了报告。
最后对计算机自适应测验软件模型的研究所存在的问题和改进的方向作
－了说明。
Computerized Adaptive Test is a new field in the modern testing research. It
is of great importance in improving testing efficiency and quality. In the
tranditionary test, all the students receive the test by the same batch of questions,
irrespective of whether the examine抯 level is high or not, As a result. The
examines of low level can抰 answer the quite difficult questions at all, but it leads
to anxiety. On the other hand, the examines of high level, those questions with low
degree of difficulty can抰 test the students?real level, it may waste energy.
CAT solved the problem in accordance with the examine, Which adapts the
examine抯 state on its own initiative. The appearance of CAT no only breaks the
paper-and-pencil test pattern which carried on from 2000 years ago and changes it
into the one by which the computer screen show and the keyboard and mouse can
be used to answer the questions, moreover, compared with the traditional. The
testing ideas have undertaken tremendous changes. That is ,it sets up a personal
test for a examine to reach an accurate measurement. Because the items choosing
is to tailor itself to the ability of the test taker, each question that the examine
answers all matches their capability. That is to say, the high-ability examine can
avoid answering the quite easy questions, but the low-ability one can avoid
answering the questions which are beyond his competence from the testing train of
though, it抯 not difficult to find out CAT抯 numerous advantages.(1) The testing is
efficiency.(2) It抯 more comparative. Because of the specific implementing and
scoring ways, CAT抯 final score can be turned into the comparative and explicable
measurement.
At present, in some developed countries, such as America, CAT has already
distinguished itself in the fields of educational test, professional measurement and
personnel matters testing, for example, American Graduate Record Examination,
Graduate for Management and Administration Entrance Test and the Entire
American Nurse National Committee License Test and so on. all of the above
ii?
/
~#i ~
\IASTER S THESrS
adapt CAT way. It抯 not difficult to find out that CAT stands for the developing
orientation and in the future education and psychological testing.
But in China, the computer software on CAT can be hardly seen, the National
College English Band four and six Test committee is getting down to the research
in this aspect. But at present there抯 no finished product software model coming
out. Therefore, to go deep into the CAT research is of important significance in
Chinese testing technology.
The first part of this article begins with analyzing the traditional testing
problems, and it puts forward the meaning and necessity of CAT, meanwhile, it
illustrates the CAT抯 inside and outside developing state.
The second part analyses the theoretical basis of CAT. The discussion centers
on JRT and the effect of the theory in the testing work, it analyses such difficulty,
difference, credibly, efficiency as important concepts in greater detail.
The third part raises the design way and implementation course on CAT抯
software model. At the same time, it lists some important arithmetic, including
making up arithmetic and sampling tactics and
效率与质量有着重要意义。传统的测验，不论考生水平高低全部要接受同一
批题目的测验，结果对水平低的考生来说，那些难度大的题目，根本无法作
答，反而引起焦虑。对高水平考生来说，那些难度小的题目，又测不出考生
的真实水平，徒然浪费精力。
计算机化自适应考试（Computer Adaptive Test，简称 CAT）解决了主动适
应考生状况的“因人施测”问题。CAT的出现不仅打破了两千多年沿袭下来的
以纸和笔作为作答工具的方式而改为计算机显示屏呈现、键盘与鼠标进行作
答的方式，而且与传统的测验相比，测验思想也发生了巨大的变革：它通过
给每一个被试建立一个个人化的测验来达到更为准确的测量，因为项目的选
择是根据被试的能力水平定身度量而成的，因而被试所做的每一个题目的难
度都是与其能力相匹配的。也就是说，水平高的被试能够避免做到相当简单
的题目，而能力低的被试能够避免做到超出其能力范围之外的题目。从其测
验思路中我们不难看出CAT考试的众多优点：（1）测验效率高；（2）可比
性强。由于特定的实施方式与记分方式，CAT最终的分数可转化为可相互比较
以及解释的量尺。
目前，在一些发达的国家如美国，CAT已在教育测验、职业测量、人事
测评等领域中大显身手，如美国研究生入学考试（Graduate Record
Examination）、工商管理类研究生入学考试（Graduate for Management and
Administration Test）以及全美护士国家委员会资格考试（Nurse National
Comittee License Test）等都已采取了CAT的方式。不难看出CAT代表了
今后教育、心理测验发展的方向与重点。
但是，在我国，有关自适应测验的计算机软件极为少见，国家大学英语
四、六级考试委员会正着手这方面的研究，但目前还没有成品的软件模型问
OS差霎自SI 砒十堂o为丫
Y’。。。·。八！u工ER S 工HESIS
世，因此，深入计算机自适应测验的研究对于我国测验技术的发展具有重要
的意义。
本文第一部分从分析传统测验的问题出发，提出了计算机自适应测验的
意义和必然性，同时，对国内外计算机自适应测验发展的状况进行了说明。。
第二部分对计算机自适应测验的理论基础进行了分析，着重讨论了项目
反应理论以及本理论在测验工作中的作用。详细地分析了难度、区分度、信
度和效度等重要的概念。
第三部分提出了计算机自适应测验软件模型的设计方法和实施过程，同
时还列出了自适应测验中的一些重要的算法，包括施测算法、组卷算法和抽
题策略等。
第四部分分析了所完成的计算机自适应测验软件模型方案设计。对其中
的功能设计，测验过程及其实现等作了详细的叙述。
第五部分对本软件的测试结果作了报告。
最后对计算机自适应测验软件模型的研究所存在的问题和改进的方向作
－了说明。
Computerized Adaptive Test is a new field in the modern testing research. It
is of great importance in improving testing efficiency and quality. In the
tranditionary test, all the students receive the test by the same batch of questions,
irrespective of whether the examine抯 level is high or not, As a result. The
examines of low level can抰 answer the quite difficult questions at all, but it leads
to anxiety. On the other hand, the examines of high level, those questions with low
degree of difficulty can抰 test the students?real level, it may waste energy.
CAT solved the problem in accordance with the examine, Which adapts the
examine抯 state on its own initiative. The appearance of CAT no only breaks the
paper-and-pencil test pattern which carried on from 2000 years ago and changes it
into the one by which the computer screen show and the keyboard and mouse can
be used to answer the questions, moreover, compared with the traditional. The
testing ideas have undertaken tremendous changes. That is ,it sets up a personal
test for a examine to reach an accurate measurement. Because the items choosing
is to tailor itself to the ability of the test taker, each question that the examine
answers all matches their capability. That is to say, the high-ability examine can
avoid answering the quite easy questions, but the low-ability one can avoid
answering the questions which are beyond his competence from the testing train of
though, it抯 not difficult to find out CAT抯 numerous advantages.(1) The testing is
efficiency.(2) It抯 more comparative. Because of the specific implementing and
scoring ways, CAT抯 final score can be turned into the comparative and explicable
measurement.
At present, in some developed countries, such as America, CAT has already
distinguished itself in the fields of educational test, professional measurement and
personnel matters testing, for example, American Graduate Record Examination,
Graduate for Management and Administration Entrance Test and the Entire
American Nurse National Committee License Test and so on. all of the above
ii?
/
~#i ~
\IASTER S THESrS
adapt CAT way. It抯 not difficult to find out that CAT stands for the developing
orientation and in the future education and psychological testing.
But in China, the computer software on CAT can be hardly seen, the National
College English Band four and six Test committee is getting down to the research
in this aspect. But at present there抯 no finished product software model coming
out. Therefore, to go deep into the CAT research is of important significance in
Chinese testing technology.
The first part of this article begins with analyzing the traditional testing
problems, and it puts forward the meaning and necessity of CAT, meanwhile, it
illustrates the CAT抯 inside and outside developing state.
The second part analyses the theoretical basis of CAT. The discussion centers
on JRT and the effect of the theory in the testing work, it analyses such difficulty,
difference, credibly, efficiency as important concepts in greater detail.
The third part raises the design way and implementation course on CAT抯
software model. At the same time, it lists some important arithmetic, including
making up arithmetic and sampling tactics and
引文
1．薛理银：《教育信息处理原理》北京师大出版社，1996。
2．漆书青 戴海琦：《项目反应理论及其应用研究》江西高校出版社，1992。
3．韩布新 车宏生：《计算机施测对心理测验的影响研究述评》《心理与教育测量》浙江：浙江教育出版社，1997。
4．王宝墉：《现代测验理论》台湾心理出版社，1983。
5．孔祥斌 胡文斌：《电脑自适应测验、解题辅导题库》《心理与教育测量》浙江教育出版社，1997。
6．淑慧 范光辉：《电脑自适性测验应用于学业性测验之成效评估》
7．张厚粲：《心理测量学的研究与发展》《心理与教育测量》浙江教育出版社，1997。
8. Weiss D J:《New horizons in testing: Latent trait test theory and computerized adaptive testing》New York: Academic Press, 1983.
9. Weiss D J:《Improving measurement quality and efficiency with computerized adaptive testing》Applied Psychology Measurement, 1982, 6:473-492.
10. Lord M F:《Applications of item response theory to practical testing problems》Hillsdale NJ: Erlbaum, 1990。
11. Cover T M, Thomas J A: 《Elements of information theory》New York: John Wiley & Sons, 1991.
12. Chang H H, Ying Z I: 《A Global Information Approach to Computerized Adaptive Testing》Applied Psychological Measurement, 1996, (2):213-219.
13. Kullback S:《Information theory and statistics》New York: Wiley, 1959.
14． Hambleton, R.K., H. Swamination, and H.J. Rogers. : 《 Fundamentals of Item Response Theory》 Newbury Park: Sage Publications. 1991．
15． Hambleton, R.K. and J.N. Zaal, eds.: 《 Advances in Educational and Psychological Testing》 Boston: Kluwer Academic Publishers. 1991．
16． Kingsbury, G.G. and A.R. Zara.: 《Procedures for selecting items for computerized adaptive tests》 Applied Measurement in Education 2: 359-375． 1989．
17． Lord,F.M: 《Applications of Item Response Theory to Practical Testing Problems》 New Jersey: Lawrence Erlbaum Associates, Publishers. 1980
18． Wainer H : 《Computerized adaptive testing: A primer》 Hillsdale NJ: Erlbraum , 1990．
19． Hambletonm R K, Swaminathan H: 《Item response theory: Principles and application》 Boston: Kluwer Nijhoff Publishing , 1985．
20． Wright B Bo Panchapakesan N A : 《procedure for sample free item analysis》 Journal of Educational and psychological Measurement 29, 23-48．
21． McKinley, R.L. and M.D. Recase.: 《Computer applications to ability testing》 Association for Educational Data Systems Journal 13: 193-203． 1980．
22． Reckase, M.D.: 《Adaptive testing: the evolution of a good idea》 Educational Measurement: Issues and Practice 8: 11-15． 1989．
23． Van der Linden, W.J. and R.K. Hambleton, eds.: 《Handbook of Modern Item Response Theory》 New York: Springer-Verlag. 1996．
24． Wainer, H.: 《Computerized Adaptive Testing》 A Primer. New Jersey: Lawrence Erlbaum Associates, Publishers. 1990．
25． Weiss, D.J..: 《New Horizons in Testing》 Latent Trait Test Theory and Computerized Adaptive Testing. New York: Academic Press. 1983
2．漆书青 戴海琦：《项目反应理论及其应用研究》江西高校出版社，1992。
3．韩布新 车宏生：《计算机施测对心理测验的影响研究述评》《心理与教育测量》浙江：浙江教育出版社，1997。
4．王宝墉：《现代测验理论》台湾心理出版社，1983。
5．孔祥斌 胡文斌：《电脑自适应测验、解题辅导题库》《心理与教育测量》浙江教育出版社，1997。
6．淑慧 范光辉：《电脑自适性测验应用于学业性测验之成效评估》
7．张厚粲：《心理测量学的研究与发展》《心理与教育测量》浙江教育出版社，1997。
8. Weiss D J:《New horizons in testing: Latent trait test theory and computerized adaptive testing》New York: Academic Press, 1983.
9. Weiss D J:《Improving measurement quality and efficiency with computerized adaptive testing》Applied Psychology Measurement, 1982, 6:473-492.
10. Lord M F:《Applications of item response theory to practical testing problems》Hillsdale NJ: Erlbaum, 1990。
11. Cover T M, Thomas J A: 《Elements of information theory》New York: John Wiley & Sons, 1991.
12. Chang H H, Ying Z I: 《A Global Information Approach to Computerized Adaptive Testing》Applied Psychological Measurement, 1996, (2):213-219.
13. Kullback S:《Information theory and statistics》New York: Wiley, 1959.
14． Hambleton, R.K., H. Swamination, and H.J. Rogers. : 《 Fundamentals of Item Response Theory》 Newbury Park: Sage Publications. 1991．
15． Hambleton, R.K. and J.N. Zaal, eds.: 《 Advances in Educational and Psychological Testing》 Boston: Kluwer Academic Publishers. 1991．
16． Kingsbury, G.G. and A.R. Zara.: 《Procedures for selecting items for computerized adaptive tests》 Applied Measurement in Education 2: 359-375． 1989．
17． Lord,F.M: 《Applications of Item Response Theory to Practical Testing Problems》 New Jersey: Lawrence Erlbaum Associates, Publishers. 1980
18． Wainer H : 《Computerized adaptive testing: A primer》 Hillsdale NJ: Erlbraum , 1990．
19． Hambletonm R K, Swaminathan H: 《Item response theory: Principles and application》 Boston: Kluwer Nijhoff Publishing , 1985．
20． Wright B Bo Panchapakesan N A : 《procedure for sample free item analysis》 Journal of Educational and psychological Measurement 29, 23-48．
21． McKinley, R.L. and M.D. Recase.: 《Computer applications to ability testing》 Association for Educational Data Systems Journal 13: 193-203． 1980．
22． Reckase, M.D.: 《Adaptive testing: the evolution of a good idea》 Educational Measurement: Issues and Practice 8: 11-15． 1989．
23． Van der Linden, W.J. and R.K. Hambleton, eds.: 《Handbook of Modern Item Response Theory》 New York: Springer-Verlag. 1996．
24． Wainer, H.: 《Computerized Adaptive Testing》 A Primer. New Jersey: Lawrence Erlbaum Associates, Publishers. 1990．
25． Weiss, D.J..: 《New Horizons in Testing》 Latent Trait Test Theory and Computerized Adaptive Testing. New York: Academic Press. 1983