Can a traditional survey reduce the bias of mobile phone data based statistics?
Conference
65th ISI World Statistics Congress 2025
Format: IPS Abstract - WSC 2025
Keywords: bias reduction,, mobile_network_data, survey
Session: IPS 934 - Integrating Mobile Network Operator Data with Official Statistics
Thursday 9 October 2 p.m. - 3:40 p.m. (Europe/Amsterdam)
Abstract
Mobile Network Operator (MNO) data has high potential for generating insights into public behaviors not just for market analysis but also for official statistics and policy-making. However, the inherent biases in MNO data, such as device-user ambiguity, multiple device usage, and varying user behaviors across demographic groups, compromise the accuracy and reliability of statistics derived from this data source. This presentation explores the possibility and usefullness of using a traditional survey to reduce these biases and enhance the quality of MNO data-based statistics. We propose an ad-hoc survey specifically designed to address the issues in MNO data by collecting detailed information on device usage, contractual relationships, and user demographics. The survey aims to determine the number of devices and SIM cards used by individuals, identify discrepancies between contract holders and actual device users, and analyze usage patterns across different demographic groups. A friendly user test of the proposed survey was conducted to gather preliminary insights and feedback. The presentation will present these findings and discuss the implications for improving the reliability of MNO-derived statistics. Additionally, complementary approaches, such as an area-based survey, will be explored as potential alternatives to further improve data coverage and quality.