我的毕业论文写完了,现在就差英文摘要部分还没定稿。我把我的摘要的中英文对照版发过来,请大家帮忙看一下有没有什么地方翻译得不对的。谢谢了啊!
摘要
Abstract
测验研究领域内寻找测量同一心理品质的两个测验形式之间分数转换关系的统计技术,叫等值。等值来源于实际工作的需要,其目的是为了使得两个不同测验形式之间的分数有可比性。
The statistical technique used to convert two different forms of tests which measure the same psychological quality into a comparable form is called equating. Equating derives from the need of practical work, and its purpose is to make two different form test scores have comparability.
等值的方法有很多种,其中在经典测验理论(CTT)下的方法主要有线性等值和等百分位等值两种。不同的等值方法会产生不同的等值结果。到底用哪种等值方法得到的结果更加精确,就成为学者们关注的问题。对此,国内外已经有过许多研究,但由于每个研究所采用的研究情境各不相同,因此结论也各不相同。
There are many methods of test equating. Among them, the two most common methods based on Classical Test Theory (CTT) are linear equating and equipercentile equating. Different methods of test equating would lead to different results. Thus, scholars are concerned about which method would lead to the most accurate result. For this, there are many researches at home and abroad. However, every research context is different, thus the conclusions are not the same.
本研究用蒙特卡洛模拟研究方法,用单组非锚测验设计,以真分数等值为依据,综合比较了各种测验难度分布条件下和各种样本容量条件下两种CTT等值方法的等值结果。
This research uses Monte Carlo method and single group design without anchor test. Based on the true score link, this paper compare the two CTT equating methods in various difficulty distributions of tests and in various sample sizes synthetically.
研究结果表明,在本研究所设情境中:
In this research context, results show as follows:
(1)线性等值的误差受测验难度分布影响较大,等百分位等值的误差几乎不受测验难度分布影响。
(1) Difficulty distributions of tests affect the error of linear equating much, while they affect the error of equipercentile equating little.
(2)线性等值的误差几乎不受样本容量的影响,等百分位等值的误差受样本容量影响较大。
(2) Sample sizes affect the error of linear equating little, while they affect the error of equipercentile equating much.
(3)不论测验难度分布如何,只要样本容量足够大,等百分位等值的效果都比线性等值更好。
(3) No matter how the difficulty distributions of tests are, equipercentile equating is better than linear equating if only the sample sizes are large enough.
本研究的结论和以往研究有一些不同之处,为此本文也进行了一些讨论。
The conclusions of this research have some differences with previous researches. They are also discussed in the paper.
关键词:测验等值;线性等值;等百分位等值;难度分布;样本容量
Key words: test equating; linear equating; equipercentile equating; difficulty distribution; sample size
|