|
Learning objectives
|
Summary
This chapter provides an overview of the key data sources for migration data, providing examples of how they could be used to collect and generate migration data and highlighting their respective strengths and limitations. As shown in the figure below, sources of migration data can be broadly grouped into three categories.
Statistical data sources include population censuses and surveys, the former of which (typically) provide universal coverage and are the most comparable across countries. Including questions on country of citizenship, country of birth, and time of arrival are key to measuring international migration. However, including questions on place of residence one and five years prior to enumeration or on past migration events for all respondents would also capture return migration or internal migration. Household surveys can be designed with several modules to capture immigration, emigration, and return migration. Censuses and surveys allow for comparisons between migrants and citizens, can be developed based on international statistical standards (e.g., UNDESA, 2022), and generally provide nationally representative samples. However, in most contexts, censuses are only conducted every 10 years and therefore unable to reflect changes during the intercensal period and statistical data sources in general are affected by under-sampling of migrants, who remain a minority group. Furthermore, in some contexts statistical data sources are not made publicly available.
Administrative data generated to support administrative processes, such as entries and exits at border posts, applications for residence permits or international protection, are not recognized as official statistics but may provide stock and flow or, if linked by a unique personal identification number (PIN), more detailed information on migrants’ profiles. However, these data are limited in that they count movements or records but not persons, they represent intended as opposed to actual migration (which means that some applicants may decide to stay for a shorter period or not make use of their permits but are still falsely counted as international migrants), permits may be renewed or extended for the same person, and dependents are sometimes counted on the same permit.
Operational data, such as IOM’s Displacement Tracking Matrix (DTM), are collected by a range of organizations to inform decisions concerning programming, responses and resource allocation associated with humanitarian action. These data are not counted as official statistics but are sometimes the most reliable migration data, for example in emergency contexts where the capacities of national statistical systems are limited or for studying forced displacement, trafficking in persons or smuggling.
Finally, innovative sources include ‘big’ data. Big data are generated automatically by the users of mobile phones, internet platforms and applications, as well as via sensors. Big data are characterised by the "3 Vs": a) volume; b) velocity, or the speed of data which are generated in (almost) real-time; and c) variety, or the complexity and unstructured information derived from the data (Hilbert, 2013). Big data sources cannot replace traditional data sources but can complement them by reducing the costs of data collection, increasing the timeliness of data, and providing data on temporary mobility, forced displacement and ‘hard-to-reach’ populations. Innovative data sources suffer from sampling bias, privacy issues, an under-developed regulatory framework, weak private sector partnerships and limited technical capacities.
None of these sources should be treated as a panacea for the current challenges to migration data and statistics. Instead, a comprehensive and reliable understanding of migration should be based on the triangulation of different data sources. This is especially true when considering that some data are more suitable for identifying certain categories of migrants while others are less suitable.