Learning Modality Choices Among Chinese Older Adults: A Classification and Regression Tree Analysis

This study employs five classification and regression tree (CART) models to explore factors associated with older adults’ learning modality choices and to derive exploratory learner profiles. A questionnaire survey of 1,147 adults aged 60 and older was conducted across eastern, central, and western China. Five binary models – face-to-face, online, blended, interest group/peer-assisted, and self-directed learning – were fitted separately, as respondents could select multiple modalities. Eight independent variables spanned demographics, socioeconomic status, and learning engagement. Pronounced class imbalance was observed (positive-class proportions: 26.3%–92.1%). Under imbalance-robust metrics, only the face-to-face model showed comparatively robust discriminatory capacity (Macro-F1 = 0.740); the apparent strength of the blended and interest group/peer-assisted models largely reflected skewed class distributions, while the online and self-directed models showed limited discriminatory power, with repeated stratified cross-validation yielding a mean Macro-F1 of 0.434 (SD = 0.019) across the five models. The splitting structure differed across modalities: living arrangement and primary daily activity opened the face-to-face and self-directed learning trees, respectively; weekly learning duration opened the online and interest group/peer-assisted trees; gender opened the blended learning tree; and community type recurred beyond the root node in all five trees, with educational attainment and age group appearing in the upper nodes of some models. As an exploratory synthesis of the splitting paths, four exploratory learner profiles were derived as interpretive ideal types rather than empirically validated categories: the digitally integrated, the traditional-social, the autonomous-independent, and the low-engagement learner. This study extends decision tree methods from older adult(s) health to educational gerontology, revealing hierarchical splitting patterns and population heterogeneity in learning modality choices. Given the cross-sectional design, the findings are associational rather than causal and provide preliminary evidence for shifting older adult education from uniform provision toward differentiated and inclusive lifelong learning.

Authors

Institutions

Publication Details

Journal
Educational Gerontology
Published
2026-11-17
DOI
https://doi.org/10.1080/03601277.2026.2760975
Primary Topic
Adult and Continuing Education Topics
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Learning Modality Choices Among Chinese Older Adults: A Classification and Regression Tree Analysis

Bing Li, Shangjing Yin
2 links
Educational Gerontology
Adult and Continuing Education Topics
article

Learning Modality Choices Among Chinese Older Adults: A Classification and Regression Tree Analysis

Bing Li, Shangjing Yin
article en

Abstract

This study employs five classification and regression tree (CART) models to explore factors associated with older adults’ learning modality choices and to derive exploratory learner profiles. A questionnaire survey of 1,147 adults aged 60 and older was conducted across eastern, central, and western China. Five binary models – face-to-face, online, blended, interest group/peer-assisted, and self-directed learning – were fitted separately, as respondents could select multiple modalities. Eight independent variables spanned demographics, socioeconomic status, and learning engagement. Pronounced class imbalance was observed (positive-class proportions: 26.3%–92.1%). Under imbalance-robust metrics, only the face-to-face model showed comparatively robust discriminatory capacity (Macro-F1 = 0.740); the apparent strength of the blended and interest group/peer-assisted models largely reflected skewed class distributions, while the online and self-directed models showed limited discriminatory power, with repeated stratified cross-validation yielding a mean Macro-F1 of 0.434 (SD = 0.019) across the five models. The splitting structure differed across modalities: living arrangement and primary daily activity opened the face-to-face and self-directed learning trees, respectively; weekly learning duration opened the online and interest group/peer-assisted trees; gender opened the blended learning tree; and community type recurred beyond the root node in all five trees, with educational attainment and age group appearing in the upper nodes of some models. As an exploratory synthesis of the splitting paths, four exploratory learner profiles were derived as interpretive ideal types rather than empirically validated categories: the digitally integrated, the traditional-social, the autonomous-independent, and the low-engagement learner. This study extends decision tree methods from older adult(s) health to educational gerontology, revealing hierarchical splitting patterns and population heterogeneity in learning modality choices. Given the cross-sectional design, the findings are associational rather than causal and provide preliminary evidence for shifting older adult education from uniform provision toward differentiated and inclusive lifelong learning.

Educational Gerontology
Open University of China (CN)
Quality education, Good health and well-being
Adult and Continuing Education Topics
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon