Findings on the cognitive causes of dyslexia are often inconsistent. We address the methodological issues that may contribute to this inconsistency. We assess the psychometric reliability of tasks widely used in perceptual theories of dyslexia. We will report data from our empirical investigation with 200 German native-speaking children in 3rd and 4th grades. Tasks include visual and auditory processing experiments, replicating previous studies: (1) rise time and frequency discrimination (Goswami et al., 2011, 2010), (2) auditory and visual temporal order judgement (Landerl & Willburger, 2010), (3) visual-spatial attention tasks (Bertoni et al., 2021), (4) coherent motion detection (Stein, 2014). In addition, we assess reading fluency and include control tasks (phonological awareness and simple and choice reaction time). Each participant undergoes two sessions — the first online (via Pavlovia.org), and the second either in the lab or online.

Preliminary data analysis, comprising 158 online and 77 in-lab datasets, reveals test-retest correlations of .34 to .45 for experimental tasks and .43 to .54 for control tasks. Recruitment continues to reach 200 datasets. We will present test-retest reliability, internal consistency, and psychometric functions of tasks. Additionally, we will investigate the association between perceptual task performance and reading ability, shedding light on the stability of effects reported in replicated studies. Insights into the feasibility of online dyslexia/reading studies with child participants will be also shared with the audience. Through discussions of methodological considerations, our goal is to enhance the reliability of dyslexia research and contribute to more coherent investigations of this complex disorder.