"Can 3D Augmented Reality Visualization help people understand Multiple Linear Regression?"

Read full text
Authors: Andrew Hill
Cited by: 0

Abstract

"Data Visualization enables people to relate abstract statistical concepts to visual representations, and prior work shows that even untrained observers can extract summary statistics from 2D scatterplots via ensemble perception. However, 3D scatterplots are often harder to interpret, despite their potential for introducing and visualizing multivariate relationships and interactions. A key limitation may be that 3D visualizations are typically displayed on 2D media. We hypothesized that presenting 3D scatterplots in true three-dimensional space would improve comprehension of basic statistical concepts. To test this, we conducted an experiment comparing participants\u2019 ability to extract information from 3D scatterplots shown on a standard 2D display or using Augmented Reality (AR) headsets. After training in Simple and Multiple Linear Regression, participants completed interpretation tasks and reported their learning experience using a modified Intrinsic Motivation Inventory. While task performance did not differ between groups for Simple Linear Regression, AR yielded a small, non-significant improvement for Multiple Regression. Notably, AR participants reported greater engagement, autonomy, and lower pressure during learning."

Fields of Study & Subjects

Data Visualization and Analytics Augmented Reality Applications Statistics Education and Methodologies

Publication details

Journal / source
IEEE DataPort
Publication date
9 March 2026
Language
EN
Document type
Dataset