An EM Algorithm for Estimating an Oral Reading Speed and Accuracy Model

TL;DR

This paper introduces a two-part Gaussian-based model for estimating oral reading speed and accuracy, utilizing an EM algorithm for parameter estimation and demonstrating its effectiveness on real data.

Contribution

It develops a novel two-part model combining speed and accuracy components with an EM algorithm for estimation, advancing assessment methods in reading research.

Findings

Model effectively captures reading speed and accuracy.

The EM algorithm provides reliable parameter estimates.

Application shows good predictive performance.

Abstract

This study proposes a two-part model that includes components for reading accuracy and reading speed. The speed component is a log-normal factor model, for which speed data are measured by reading time for each sentence being assessed. The accuracy component is a binomial-count factor model, where the accuracy data are measured by the number of correctly read words in each sentence. Both underlying latent components are assumed to be Gaussian in nature. In this paper, the theoretical properties of the proposed model are developed and an Monte Carlo EM algorithm for model fitting is outlined. The predictive power of the model is illustrated in a real data application.