词汇复杂度三种测量方法的对比分析
发布时间：2018-09-01 16:03
【摘要】：词汇评估是二语习得研究的一个重要领域。词汇丰富性是二语学习者口笔语词汇产出能力的重要评估类别，该类别包含的指标及其测量的准确性是近几十年二语词汇评估中的热点问题。词汇丰富性的测量指标包括词汇多样性、词汇复杂度、词汇密度和词汇独特性等。其中，，二语学习者词汇复杂度的测量方法是最引起争议的。前期研究中二语学习者词汇复杂度的三种测量方法—词频概貌（Lexical Frequency Profile, LFP），P_Lex，Advanced D都声称自己是最好的，但哪一种最佳还没有达成一致。因此，本研究旨在通过比较二语学习者词汇复杂度的三种测量方法，试图找出最佳的测量方法,并为词汇研究测量提供新路子。具体研究以下问题： 1）从信度方面看，二语学习者词汇复杂度的三种测量方法哪种最佳？ 2）从效度方面看，二语学习者词汇复杂度的三种测量方法哪种最佳？ 为分析三种测量方法的信度，本研究使用50名英语专业三年级学生的两套同一题材的随堂限时作文，第二篇作文与第一篇作文相隔两周。 本研究从构念效度和共现效度两个方面分析三种测量方法的效度。二语学习者词汇复杂度的构念效度的研究涉及到测量文本长度和高级词的变化性对三种测量方法的影响。本研究通过对来自于TEM-8考试作文中的三篇作文进行描述性分析检验三种测量方法的构念效度。三篇作文对应的分数分别是18分，13分，9分，它们依次称作高、中、低水平组。基于这三篇作文得到的计算机模拟数据用于测量高级词的变化性对三种测量方法的影响。在第一种模拟中，高级标记总数保持不变，高级词型数依次减少，在第二种模拟中，高级标记数随高级词型数依次下降。 为分析三种测量方法的共现效度，本研究从英语专业八级考试作文中随机选取80篇作文。本研究中学习者八级作文成绩(满分为20分)分布在8～18分之间，其中分布最集中的两个分值是12和13分，我们按这两个分值将作文成绩分为两个等级:8～12分的学习者为低水平组，共42人；13～18分的学习者为高水平组，共38人。 本研究的主要发现如下： 信度方面，皮尔逊相关分析发现,两组作文中二语学习者词汇复杂度的三种测量方法均达到显著的相关,但是相关度不高(LFP：r1=.327, p1=.020㩳.05；P_Lex：r2=.308, p2=.030㩳.05；Advanced D：r3=.441, p3=.001㩳.05)。配对样本t检验表明,两组作文的词汇复杂度有显著差异(LFP：t1=4.804, p1=.000㩳.05；P_Lex：t2=8.837, p2=.000㩳.05；Advanced D：t3=-2.742,p3=.008㩳.05)。 构念效度方面，在测量文本长度对三种测量方法的影响时，LFP，P_Lex和Advanced D对文本长度的最低要求分别是200个词，120个词和120个词。但是Advanced D要略好于P_Lex，因为在中低水平组作文中Advanced D获得较为稳定的结果是60个词，而P_Lex是120个词。在测量高级词的变化性对三种测量方法的影响时，三篇不同水平作文的LFP和P_Lex的词汇复杂度值在第一种模拟中几乎保持不变，而在第二种模拟中不断下降波动。然而，Advanced D在两种模拟中的词汇复杂度值均呈现不同程度的变化。这表明Advanced D在测量高级词的变化性对词汇复杂度的影响时要好于P_Lex和LFP。 共现效度方面，相关分析发现三种测量方法与学习者英语作文质量呈现微弱正相关(LFP：r1=.248, p1=.027㩳.05；P_Lex：r2=.253, p2=.024㩳.05；Advanced D：r3=.257, p3=.021㩳.05)，词汇复杂度和词汇多样性呈现正相关(LFP：r1=.332, p1=.003㩳.05；P_Lex：r2=.236, p2=.035㩳.05；Advanced D：r3=.340, p3=.002㩳.05)，且Advanced D和LFP略高于P_Lex。 总之，Advanced D和P_Lex在控制文本长度对二语学习者词汇复杂度影响方面要比LFP更稳定。而在测量高级词的变化性对词汇复杂度影响方面，Advanced D要好于P_Lex和LFP。综上所述，无论从信度看还是效度看，AdvancedD是三种方法中最佳的词汇复杂度测量方法。本研究不仅在方法论上对二语学习者词汇复杂度测量做出了贡献，而且对词汇评估也有重要意义。
[Abstract]:Vocabulary richness is an important category in the study of second language acquisition (SLA). The indicators and the accuracy of vocabulary richness measurements have been a hot issue in the field of SLA vocabulary assessment in recent decades. Among them, the measurement of L2 learners'lexical complexity is the most controversial. In the previous study, three methods of measuring L2 learners' lexical complexity, namely, Lexical Frequency Profile (LFP), P_Lex, and Advanced D, all claimed to be the best, but which one was the best, were not yet available. Consequently, this study aims to find out the best way to measure L2 learners'lexical complexity by comparing the three measurements and to provide a new way for lexical research and measurement.
1) from the perspective of reliability, what are the three best ways to measure the lexical complexity of two language learners?
2) from the perspective of validity, what are the three best ways to measure the lexical complexity of two language learners?
In order to analyze the reliability of the three measurement methods, 50 third-year English majors were asked to write two sets of time-limited compositions on the same topic in class. The second one was two weeks apart from the first one.
This study examines the validity of the three measurement methods in terms of constructive validity and co-occurrence validity. The constructive validity of L2 learners'lexical complexity involves measuring the effects of text length and the variability of advanced words on the three measurement methods. The construct validity of the three methods was analyzed and tested. The corresponding scores of the three compositions were 18, 13, and 9. They were called high, medium and low-level groups in turn. In the second simulation, the number of advanced markers decreases with the number of advanced morphologies.
In order to analyze the co-occurrence validity of the three measurement methods, 80 compositions were randomly selected from the compositions of CET-8. In this study, the scores of CET-8 compositions (20 full marks) ranged from 8 to 18. The two most concentrated scores were 12 and 13. We divided the compositions into two grades according to the two scores. A total of 8~12 learners were 42 in the low level group, and 13~18 in the high level group, 38 in total.
The main findings of this study are as follows:
In terms of reliability, Pearson correlation analysis found that there was a significant correlation between the three measurement methods of LFP: R1 =.327, P1 =.020?.05; P_Lex: R2 =.308, P2 =.030?.05; Advanced D: R3 =.441, P3 =.001?.05). The paired sample t test showed that the lexical complexity of the two groups'compositions was not significant. There were significant differences in impurity (LFP: T1 = 4.804, P1 =.000?.05; P_Lex: T2 = 8.837, P2 =.000?.05; Advanced D: T3 =-2.742, P3 =.008?.05).
In terms of constructive validity, LFP, P_Lex and Advanced D required 200 words, 120 words and 120 words respectively when measuring the effect of text length on the three measurement methods. The lexical complexity values of LFP and P_Lex in three compositions with different proficiency levels remained almost unchanged in the first simulation, but declined and fluctuated in the second simulation. However, the lexical complexity values of Advanced D showed varying degrees of variation in both simulations. This indicates that Advanced D is better than P_Lex and LFP in measuring the influence of the variability of advanced words on lexical complexity.
In terms of co-occurrence validity, the correlation analysis found that there was a weak positive correlation between the three measurement methods and the quality of English writing (LFP: R1 =.248, P1 =.027?.05; P_Lex: R2 =.253, P2 =.024?.05; Advanced D: R3 =.257, P3 =.021?.05), and lexical complexity and lexical diversity were positively correlated (LFP: R1 =.332, P1 =.003?.05; P_Lex: R2 =.236, P2 =.024?.05). =.035?.05; Advanced D:r3=.340, p3=.002?.05, and Advanced D and LFP are slightly higher than P_Lex..
In summary, Advanced D and P_Lex are more stable than LFP in controlling the effect of text length on L2 learners'lexical complexity, while Advanced D is better than P_Lex and LFP in measuring the effect of the variability of advanced words on lexical complexity. This study not only contributes to the measurement of L2 learners'lexical complexity in terms of methodology, but also plays an important role in lexical assessment.
【学位授予单位】：南京工业大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：H09
本文编号：2217681
[Abstract]:Vocabulary richness is an important category in the study of second language acquisition (SLA). The indicators and the accuracy of vocabulary richness measurements have been a hot issue in the field of SLA vocabulary assessment in recent decades. Among them, the measurement of L2 learners'lexical complexity is the most controversial. In the previous study, three methods of measuring L2 learners' lexical complexity, namely, Lexical Frequency Profile (LFP), P_Lex, and Advanced D, all claimed to be the best, but which one was the best, were not yet available. Consequently, this study aims to find out the best way to measure L2 learners'lexical complexity by comparing the three measurements and to provide a new way for lexical research and measurement.
1) from the perspective of reliability, what are the three best ways to measure the lexical complexity of two language learners?
2) from the perspective of validity, what are the three best ways to measure the lexical complexity of two language learners?
In order to analyze the reliability of the three measurement methods, 50 third-year English majors were asked to write two sets of time-limited compositions on the same topic in class. The second one was two weeks apart from the first one.
This study examines the validity of the three measurement methods in terms of constructive validity and co-occurrence validity. The constructive validity of L2 learners'lexical complexity involves measuring the effects of text length and the variability of advanced words on the three measurement methods. The construct validity of the three methods was analyzed and tested. The corresponding scores of the three compositions were 18, 13, and 9. They were called high, medium and low-level groups in turn. In the second simulation, the number of advanced markers decreases with the number of advanced morphologies.
In order to analyze the co-occurrence validity of the three measurement methods, 80 compositions were randomly selected from the compositions of CET-8. In this study, the scores of CET-8 compositions (20 full marks) ranged from 8 to 18. The two most concentrated scores were 12 and 13. We divided the compositions into two grades according to the two scores. A total of 8~12 learners were 42 in the low level group, and 13~18 in the high level group, 38 in total.
The main findings of this study are as follows:
In terms of reliability, Pearson correlation analysis found that there was a significant correlation between the three measurement methods of LFP: R1 =.327, P1 =.020?.05; P_Lex: R2 =.308, P2 =.030?.05; Advanced D: R3 =.441, P3 =.001?.05). The paired sample t test showed that the lexical complexity of the two groups'compositions was not significant. There were significant differences in impurity (LFP: T1 = 4.804, P1 =.000?.05; P_Lex: T2 = 8.837, P2 =.000?.05; Advanced D: T3 =-2.742, P3 =.008?.05).
In terms of constructive validity, LFP, P_Lex and Advanced D required 200 words, 120 words and 120 words respectively when measuring the effect of text length on the three measurement methods. The lexical complexity values of LFP and P_Lex in three compositions with different proficiency levels remained almost unchanged in the first simulation, but declined and fluctuated in the second simulation. However, the lexical complexity values of Advanced D showed varying degrees of variation in both simulations. This indicates that Advanced D is better than P_Lex and LFP in measuring the influence of the variability of advanced words on lexical complexity.
In terms of co-occurrence validity, the correlation analysis found that there was a weak positive correlation between the three measurement methods and the quality of English writing (LFP: R1 =.248, P1 =.027?.05; P_Lex: R2 =.253, P2 =.024?.05; Advanced D: R3 =.257, P3 =.021?.05), and lexical complexity and lexical diversity were positively correlated (LFP: R1 =.332, P1 =.003?.05; P_Lex: R2 =.236, P2 =.024?.05). =.035?.05; Advanced D:r3=.340, p3=.002?.05, and Advanced D and LFP are slightly higher than P_Lex..
In summary, Advanced D and P_Lex are more stable than LFP in controlling the effect of text length on L2 learners'lexical complexity, while Advanced D is better than P_Lex and LFP in measuring the effect of the variability of advanced words on lexical complexity. This study not only contributes to the measurement of L2 learners'lexical complexity in terms of methodology, but also plays an important role in lexical assessment.
【学位授予单位】：南京工业大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：H09
【相似文献】
相关期刊论文 前10条
1 刘茜;徐建平;许诺;;人事选拔中作假的内涵及测量方法[J];心理科学进展;2013年02期
2 史高岩;李仁军;张燕;张冬冬;刘金同;;校园欺负行为研究进展[J];精神医学杂志;2010年01期
3 史玲玲;;成人依恋的研究及其进展[J];科协论坛(下半月);2010年08期
4 兰生明;;内隐攻击性的测量方法概述[J];新西部(理论版);2012年04期
5 王健;王丽娜;孟庆跃;;主观幸福感测量方法及其影响因素研究[J];中国社会医学杂志;2008年04期
6 闫云霞;王启成;李丽涛;;学业拖延测量的研究综述[J];社会心理科学;2012年02期
7 孙湖!163357;起跑器安装方法小改进[J];中国学校体育;2000年04期
8 丁爱华;;让思维转个弯[J];高中数理化(高二版);2007年05期
9 陆庆五;;人类学测量方法的重大革新[J];化石;1993年04期
10 亚·帕·莫尔卓娃;我怎样在一年级和二年级里教长度和重量[J];江苏教育;1953年02期
相关会议论文 前10条
1 王占良;;涉及参数测量方法的充分公开问题探讨[A];2013年中华全国专利代理人协会年会暨第四届知识产权论坛论文汇编第二部分[C];2013年
2 张志广;周彦博;徐建一;;微循环血流速度测量方法的研究[A];首届中国仪器仪表学会医疗仪器分会学术会议论文集[C];1993年
3 徐燕新;吴仁祥;;悬高精确测量方法[A];矿山建设工程技术新进展——2008全国矿山建设学术会议文集（下）[C];2008年
4 胡清国;;高等级公路放线测量方法及分析[A];全国城市公路学会第二十三次学术年会论文集[C];2014年
5 高指林;胡云峰;徐玉强;杨啸;韩晓康;;大型钢结构制造中的测量方法浅析[A];'2012中国钢结构行业大会论文集[C];2012年
6 孙灵霞;李炬;魏肯堂;;环内止口间隙宽度的ICT精确测量方法[A];中国工程物理研究院科技年报（2000）[C];2000年
7 王者福;付岚冰;张蓬洲;;使用双样品的电子自旋共振定量测量方法[A];第七届全国波谱学学术会议论文摘要集[C];1992年
8 戴景民;王强;辛春锁;;接触式多参数动态热物性测量方法综述[A];2008中国仪器仪表与测控技术进展大会论文集（Ⅲ）[C];2008年
9 杨德顺;魏爱军;;信号源的复反射系数Γ_g的两种测量方法[A];1989年全国微波会议论文集（上）[C];1989年
10 张登攀;朱红丽;;现代生产系统中的网格化集成测量方法[A];第八届全国信息获取与处理学术会议论文集[C];2010年
相关硕士学位论文 前7条
1 滕奉涛;网络应用层时延测量方法比较研究[D];华中科技大学;2011年
2 赵汝芹;服务质量测量方法及其应用研究[D];大连海事大学;2007年
3 王党树;转子径向振动测量方法的研究[D];西安理工大学;2007年
4 侯正清;高速公路行驶车辆信息测量方法的研究[D];哈尔滨工业大学;2009年
5 李雪莲;词汇复杂度三种测量方法的对比分析[D];南京工业大学;2013年
6 许晓东;基于机器视觉的发动机凸轮轴三维测量方法研究[D];上海交通大学;2008年
7 潘小林;生物质灰“除尘特性”及测量方法的研究[D];华北电力大学;2014年
本文编号：2217681
本文链接：https://www.wllwen.com/wenyilunwen/yuyanxuelw/2217681.html
