학술논문
Lollipops Help Align Visual and Statistical Fit Estimates in Scatterplots With Nonlinear Models
Document Type
Periodical
Author
Source
IEEE Transactions on Visualization and Computer Graphics IEEE Trans. Visual. Comput. Graphics Visualization and Computer Graphics, IEEE Transactions on. 29(7):3436-3440 Jul, 2023
Subject
Language
ISSN
1077-2626
1941-0506
2160-9306
1941-0506
2160-9306
Abstract
Scatterplots overlayed with a nonlinear model enable visual estimation of model-data fit. Although statistical fit is calculated using vertical distances, viewers’ subjective fit is often based on shortest distances. Our results suggest that adding vertical lines (“lollipops”) supports more accurate fit estimation in the steep area of model curves (https://osf.io/fybx5/).