# Assessing mobile instant messenger networks with donated data

## TL;DR

This study uses donated data to analyze the structure of mobile instant messenger networks like WhatsApp, revealing key patterns in user connections.

## Contribution

The paper introduces data donation from a probability-based panel to study mobile instant messenger networks, providing the first national-level analysis of WhatsApp usage patterns.

## Key findings

- The degree distribution of contacts in WhatsApp follows a log-normal pattern.
- Group membership distribution is best modeled by an exponential distribution.
- Predictors of degree from extended network literature largely replicate in this dataset.

## Abstract

Despite the increasing popularity and societal relevance of mobile instant messenger services (MIMSs) such as WhatsApp, Telegram and Signal, the vast majority of the research on social media is focused on the “traditional” social media platforms such as Twitter (X) and Facebook, mostly due to the difficulty of accessing MIMSs data. Consequently, little is known about even the most basic topological features of societal-scale instant messenger networks. To overcome this knowledge gap, we employ the innovative approach of data donation by respondents of a high-quality probability-based panel in the Netherlands (the LISS panel) to collect user data while preserving their privacy. Focusing on WhatsApp as the most popular MIMS, this study collects the first measurement of MIMS usage on a national probability sample, focusing on degree distributions as a key feature of network topology. We find that the degree distribution of contacts is best approximated by a log-normal distribution, while the distribution of group membership is best approximated by the exponential distribution. At the individual level, we find that predictors of degree derived from the literature on extended networks mostly replicate.

The online version contains supplementary material available at 10.1007/s13278-025-01550-8.