Science Is Not a Golden Rule: A Teacher’s Journey into the Replicability Crisis
📐

Science Is Not a Golden Rule: A Teacher’s Journey into the Replicability Crisis

March 1, 2026·
Anna Yi Leung
Anna Yi Leung
· 2 min read
Cover photo image credit: Unsplash
blog

When I was studying education at university, I strongly believed in one thing: scientific research backed with empirical evidence was trustworthy. It must be.

If a paper contained experiments, data, and statistical analyses, that was scientific evidence. At most, I would check which university had published the study. The more prestigious the laboratory, the more credible it should be, right? And these education theories with scientific backing? Teachers should be able to confidently use them as guidelines when designing instructional methods, right?

Later, I went to teach in primary schools in Hong Kong. I taught the Chinese language.

In Hong Kong, many schools collaborate with universities on research projects. This is, in fact, a good thing: students receive free resources and training, schools obtain professional support, and teachers feel that they are standing at the forefront of science, contributing to research.

University research funding is generally invested in lower primary levels (Grades 1 and 2), as well as students with special educational needs (such as children with developmental dyslexia or language disorders).

When I first entered the profession, the school assigned me to teach Primary 1 Chinese. At the beginning of the semester, we used a literacy curriculum developed by a local university research team. This curriculum adopted a “concentrated character-learning approach” (Notes: in the Chinese language, a Chinese word is in the form of a character. A character that has a meaning attached to it is called a “word”.): characters sharing the same components (for example, the same semantic or phonetic radicals) were grouped together and taught systematically and intensively. In theory, this approach, concentrated character learning, should facilitate character acquisition as children can learn a large batch of semantically or phonologically-related characters in one go.

But gradually, I began to feel confused. My students became more familiar with the meanings or pronunciations of these Chinese characters’ components. However, from what I observed, concentrated learning of characters that share the same radicals did not necessarily help children differentiate and encode these orthographically / semantically / phonologically similar characters. Learning batches of characters with the same component together actually created interference and confused the meanings of those words.

How scientific, exactly, were these scientific methods? ……

Check out the full text on my Substack: Pacing Between Classroom and Lab

Anna Yi Leung
Authors
Research Scientist in Psycholinguistics and Metascience
I am a doctoral researcher in psycholinguistics at the Ludwig Maximilian University of Munich, formerly trained and served as a Chinese language teacher in Hong Kong primary and secondary schools. My research revolves around the cognitive mechanisms underlying reading development and the methodology used to identify subtypes of developmental dyslexia. I also build infrastructures based on metascientific principles to help synthesise research findings in psychological sciences. I am committed to bridging the gap between scientific research and the actual practice in the educational and clinical settings. My goal is to translate cross-disciplinary knowledge for researchers, educators, pracitioners, and parents by connecting insights across fields.