Measuring spelling skills: A meta-analysis of the convergence of dictation and error detection tasks

Abstract

Accurate spelling is essential for academic and professional achievement. Developing cost-effective assessments to provide early support for children experiencing difficulties with spelling is crucial. In the past, dictations have been the primary method for measuring spelling skills. However, given that reading and spelling rely on shared orthographic knowledge, detecting spelling errors might serve as a more efficient assessment approach. Recently, passive formats have gained prominence such as error identification tests or multiple-choice formats, which can be administered digitally. This study aimed to examine the relationship between active and passive methods of spelling assessment in a meta-analysis while examining potential moderators. A multilevel meta-analysis was conducted, utilizing 27 studies with 69 samples and 113 correlations. A strong positive overall correlation of $\bar{r}$ = .77, 95 % CI [.74, .80] was found. Stimulus material and various forms of active and passive methods for assessing spelling skills significantly influenced the relationship as moderators, whereas grade level had no impact. No evidence of publication bias was detected. The positive relationship between active and passive methods of spelling assessment supports their interchangeable use in practice. As the number of primary studies increases, future meta-analyses could explore further moderators that may influence the strong overall correlation.