Work

Learning with erroneous visualizations modulates retention depending on perceptual richness and test type

Year: 2025

Type: article

Abstract: BACKGROUND: Learners are increasingly subjected to inaccurate visualizations generated using artificial intelligence. Coincidentally, some learning strategies purposely let learners engage with erroneous content. Instructional visualizations can be created in varying levels of perceptual richness (or realism). We conducted the present experiment to assess whether testing using erroneous examples can be a viable strategy for learning tasks depending on their perceptual richness. METHODS: The two factors of testing (with an error-spotting task vs. without) and realism (schematic vs. realistic) were assessed using two retention tests. RESULTS: Testing with erroneous examples was detrimental when combined with schematic visualizations, but helpful when applied to realistic imagery in one of the tests. CONCLUSION: Studying with erroneous examples appears to require a certain level of realism to be an effective method. Exposing learners to inaccurate visualizations such as those generated by artificial intelligence tools may not be problematic as long as this fact is disclosed.

Source: Trends in Neuroscience and Education

DOI: 10.1016/j.tine.2025.100256

Authors: Theresa Dechamps, Alexander Skulmowski

Institution: Karlsruhe University of Education

Language: English

Topic: Visual and Cognitive Learning Processes

Subfield: Experimental and Cognitive Psychology

Field: Psychology

Domain: Social Sciences

SDG: Quality education

Open Access status: hybrid

Estimated APC paid: $3,880