Thirty Years of The Network Scale-up Method.

Area of Science:

• Social Sciences
• Epidemiology
• Network Analysis

Background:

• Estimating the size of hard-to-reach populations is a significant challenge across various disciplines.
• The Network Scale-up Method (NSUM) offers a novel approach using respondent-gathered network data.

Purpose of the Study:

• To provide an in-depth analysis of Aggregated Relational Data (ARD) properties and collection techniques.
• To comprehensively review and compare different NSUM estimators, including their assumptions and practical applications.
• To identify open problems and future research directions in hidden population size estimation.

Main Methods:

• Analysis of Aggregated Relational Data (ARD) properties and data collection methods.
• Comprehensive review of Network Scale-up Method (NSUM) estimators: direct, maximum likelihood, and Bayesian.
• Application and comparative performance analysis of various NSUM models on a canonical dataset.

Main Results:

• Detailed examination of ARD data characteristics and collection strategies.
• Comparative evaluation of different NSUM estimation techniques, highlighting their strengths and weaknesses.
• Empirical comparison of model performance on a specific dataset, with findings in supplementary materials.

Conclusions:

• The study provides a thorough overview of NSUM methodology for estimating hidden populations.
• It offers practical insights into choosing and implementing appropriate estimation methods.
• Identifies key areas for future research to advance the field of network-based population estimation.