A Study on the Factors Influencing Older Adults’ Participation in and Frequency of Volunteering: An Analysis Using the Zero-inflated Negative Binomial Regression

Abstract

Volunteer activities among older adults play a crucial role not only in contributing to the public good but also in enhancing their physical and mental health and fostering social relationships. However, the volunteer participation rate of older adults in South Korea remains at a relatively low level. To promote wider participation, it is essential to build a deeper understanding of the phenomenon. This study examines the frequency of volunteer activity among older adults as the dependent variable and compares the results of the Negative Binomial Regression model with those of the Zero-Inflated Negative Binomial Regression model, in order to highlight the importance of applying an analytical method that reflects the data structure. The results show that the standard Negative Binomial model fails to adequately account for the excess zero values—representing non-participation—which may lead to inaccurate estimation of explanatory variables. In contrast, the Zero-Inflated Negative Binomial model, which is better suited to data with an excessive number of zeros, yields significantly different results. Specifically, only educational attainment (positive), housing status (renter), depressive symptoms (negative), and institutional trust (negative) were found to significantly influence the intensity of volunteering among older adults in the zero-inflated model. These findings suggest that, in order to achieve an accurate understanding of older adults’ volunteer behavior, it is crucial to apply analytical approaches that align with the characteristics of the data. This study underscores that the appropriate choice and application of statistical models are essential tasks in the knowledge-building process.