Elliott Mark, Buttery Paula
University of Cambridge, Cambridge, UK.
Educ Psychol Meas. 2022 Oct;82(5):989-1019. doi: 10.1177/00131644211046253. Epub 2021 Sep 24.
We investigate two non-iterative estimation procedures for Rasch models, the pair-wise estimation procedure (PAIR) and the Eigenvector method (EVM), and identify theoretical issues with EVM for rating scale model (RSM) threshold estimation. We develop a new procedure to resolve these issues-the conditional pairwise adjacent thresholds procedure (CPAT)-and test the methods using a large number of simulated datasets to compare the estimates against known generating parameters. We find support for our hypotheses, in particular that EVM threshold estimates suffer from theoretical issues which lead to biased estimates and that CPAT represents a means of resolving these issues. These findings are both statistically significant ( < .001) and of a large effect size. We conclude that CPAT deserves serious consideration as a conditional, computationally efficient approach to Rasch parameter estimation for the RSM. CPAT has particular potential for use in contexts where computational load may be an issue, such as systems with multiple online algorithms and large test banks with sparse data designs.
我们研究了用于Rasch模型的两种非迭代估计程序，即成对估计程序（PAIR）和特征向量法（EVM），并确定了EVM在评分量表模型（RSM）阈值估计方面的理论问题。我们开发了一种新程序来解决这些问题——条件成对相邻阈值程序（CPAT），并使用大量模拟数据集对这些方法进行测试，以将估计值与已知的生成参数进行比较。我们的假设得到了支持，特别是EVM阈值估计存在理论问题，会导致估计有偏差，而CPAT是解决这些问题的一种方法。这些发现具有统计学显著性（<0.001）且效应量很大。我们得出结论，CPAT作为一种用于RSM的Rasch参数估计的条件性、计算效率高的方法，值得认真考虑。CPAT在计算负荷可能成为问题的情况下具有特别的应用潜力，例如具有多个在线算法的系统以及具有稀疏数据设计的大型题库。
