Rasch模型在高中英語試題分析與題庫建設中的應用研究
發(fā)布時間：2018-07-15 21:16
【摘要】：本文研究的目的在于探討大規(guī)�？荚囍性囶}質(zhì)量的分析的途徑與方法及高質(zhì)量題庫的初步建立。由于大規(guī)模選拔性考試對教育部門和考生個體具有重大的意義，所以對其進行科學的質(zhì)量分析尤為重要。通過對試題進行質(zhì)量分析，可以了解試題能否最大限度地表達不同能力水平學生的能力信息，并保證一定程度的公平和公正。而題庫的建立對學校教育評價體系有極大的意義。通過建立題庫，可以增加學校教育評價的診斷性功能，提高學生學習效率，并從根本上減輕學生的學業(yè)負擔和考試負擔。 本文主要應用Rasch測量理論對試題質(zhì)量進行分析。Rasch模型是對題目難度和學生能力同時進行估計的一種概率模型。它根據(jù)原始數(shù)據(jù)建立學生和測驗題目之間的線性測量結(jié)構(gòu)，得出學生與學生、測驗題目與題目以及學生與測驗題目之間的關(guān)系，并把學生能力水平和題目的難度水平放在同一個等距量尺上進行比較，可以直觀地看出學生的能力水平和題目的難度水平之間的差異，二者的差異越小，，測驗目標越明確。這樣，可以根據(jù)試題總體呈現(xiàn)的情況及每道題呈現(xiàn)的情況對試題質(zhì)量進行分析。并依此為基礎(chǔ)，結(jié)合CAT選題原則及策略，初步建立題庫。本文以2014年貴州省高中第一次模擬考試中1812名考生的答題情況為例，使用Rasch模型分析軟件對其進行分析。研究結(jié)果表明： 1）本次考試試題題目難度分布較好；試題能充分反應能力水平中等的考生的信息量，但不能充分反應能力水平較高的學生的信息。 2）基于Rasch模型能得到對題目客觀的分析數(shù)據(jù)，便于篩選出高質(zhì)量題目以建立題庫。 3）Rasch模型能使建立基層教育部門建立高質(zhì)量題庫成為可能，并能以此為基礎(chǔ)，針對不同考試目的或功能，進行科學組卷。
[Abstract]:The purpose of this paper is to explore the ways and methods of analyzing the quality of test questions in large scale examinations and the preliminary establishment of high quality question bank. Because of the great significance of the large-scale selection examination to the education department and the examinee individual, it is very important to carry on the scientific quality analysis to it. By analyzing the quality of the test questions, we can find out whether the test questions can express the ability information of students with different ability levels to the maximum extent, and ensure a certain degree of fairness and justice. The establishment of question bank is of great significance to the evaluation system of school education. By setting up the question bank, the diagnostic function of school education evaluation can be increased, the study efficiency of students can be improved, and the burden of study and examination can be lightened fundamentally. In this paper, Rasch measurement theory is used to analyze the quality of test questions. Rasch model is a probabilistic model to estimate the difficulty of the test and the ability of the students at the same time. Based on the original data, the linear measurement structure between students and test questions is established, and the relationship between students and students, between test questions and questions, and between students and test questions is obtained. By comparing the students' ability level with the difficulty level of the subject on the same equidistant scale, the difference between the students' ability level and the difficulty level of the question can be seen intuitively. The smaller the difference between the two levels, the clearer the test goal. In this way, you can analyze the quality of the test questions according to the overall presentation and the presentation of each question. And based on this, combined with cat selection principles and strategies, the preliminary establishment of the question bank. This paper takes the results of 1812 examinees in the first simulated examination of Guizhou Province in 2014 as an example and uses Rasch model analysis software to analyze them. The results show that: 1) the difficulty distribution of the test questions is better, the test questions can fully reflect the information content of the candidates with average ability, But they can not fully respond to the higher level of students' information. 2) based on Rasch model, we can get objective analysis data on the topic. The Rasch model can make it possible to set up a high quality question bank for the basic education department, and can be used as a basis for scientific composition according to the different purpose or function of the examination.
【學位授予單位】：貴州師范大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：B848.2
本文編號：2125408
[Abstract]:The purpose of this paper is to explore the ways and methods of analyzing the quality of test questions in large scale examinations and the preliminary establishment of high quality question bank. Because of the great significance of the large-scale selection examination to the education department and the examinee individual, it is very important to carry on the scientific quality analysis to it. By analyzing the quality of the test questions, we can find out whether the test questions can express the ability information of students with different ability levels to the maximum extent, and ensure a certain degree of fairness and justice. The establishment of question bank is of great significance to the evaluation system of school education. By setting up the question bank, the diagnostic function of school education evaluation can be increased, the study efficiency of students can be improved, and the burden of study and examination can be lightened fundamentally. In this paper, Rasch measurement theory is used to analyze the quality of test questions. Rasch model is a probabilistic model to estimate the difficulty of the test and the ability of the students at the same time. Based on the original data, the linear measurement structure between students and test questions is established, and the relationship between students and students, between test questions and questions, and between students and test questions is obtained. By comparing the students' ability level with the difficulty level of the subject on the same equidistant scale, the difference between the students' ability level and the difficulty level of the question can be seen intuitively. The smaller the difference between the two levels, the clearer the test goal. In this way, you can analyze the quality of the test questions according to the overall presentation and the presentation of each question. And based on this, combined with cat selection principles and strategies, the preliminary establishment of the question bank. This paper takes the results of 1812 examinees in the first simulated examination of Guizhou Province in 2014 as an example and uses Rasch model analysis software to analyze them. The results show that: 1) the difficulty distribution of the test questions is better, the test questions can fully reflect the information content of the candidates with average ability, But they can not fully respond to the higher level of students' information. 2) based on Rasch model, we can get objective analysis data on the topic. The Rasch model can make it possible to set up a high quality question bank for the basic education department, and can be used as a basis for scientific composition according to the different purpose or function of the examination.
【學位授予單位】：貴州師范大學
【學位級別】：碩士
【學位授予年份】：2014
【分類號】：B848.2
【參考文獻】
相關(guān)期刊論文 前10條
1 王曉華;文劍冰;;項目反應理論在教育考試命題質(zhì)量評價中的應用[J];教育科學;2010年03期
2 俞曉琳;項目反應理論與經(jīng)典測驗理論之比較[J];南京師大學報(社會科學版);1998年04期
3 趙守盈;何妃霞;陳維;羅杰;關(guān)丹丹;;Rasch模型在研究生入學考試質(zhì)量分析中的應用[J];教育研究;2012年06期
4 沐守寬;;CTT與IRT測量原理之比較[J];上海師范大學學報(基礎(chǔ)教育版);2006年04期
5 何蓮珍;張潔;;多層面Rasch模型下大學英語四、六級考試口語考試(CET-SET)信度研究[J];現(xiàn)代外語;2008年04期
6 J.A.Keats,陳富國;Rasch的測驗理論[J];心理學報;1990年03期
7 晏子;;心理科學領(lǐng)域內(nèi)的客觀測量——Rasch模型之特點及發(fā)展趨勢[J];心理科學進展;2010年08期
8 朱寧寧,張厚粲;CTT與IRT方法對人格測驗結(jié)果處理的比較研究[J];心理學探新;2003年03期
9 王蕾;;Rash客觀等距測量在PISA中國試測研究中的實踐[J];心理學探新;2007年04期
10 羅冠中;;Rasch模型及其發(fā)展[J];教育研究與實驗;1992年02期
本文編號：2125408
本文鏈接：http://sikaile.net/shekelunwen/xinlixingwei/2125408.html
教材專著
