2025 Theses Doctoral

Assessing Estimation Bias and Generalizability with Snowball Samples: Insights from Simulation Studies

Ye, Yingxin

Snowball sampling is a non-probability sampling technique that leverages existing participants' social networks to identify and recruit additional research subjects. It is particularly effective for accessing hard-to-reach populations and capturing social relations.

Early-stage researchers often adopt snowball sampling as a convenience sampling tool for quantitative studies via social media and online surveys. However, the fundamental methodological problems of non-representativeness and dependent observation in snowball samples are often overlooked.

The present research aims to conduct a thorough examination of snowball sampling by exploring how sampling designs and the underlying population network structures affect estimation bias and sample generalizability through two integrated simulation studies.

By systematically investigating the complex interactions between sampling strategies and population network properties, this research is meant to provide empirical insights into the methodological challenges of snowball sampling and to enable the development of nuanced recommendations for its appropriate application in social science research.