Like any other data source, using MPD in producing official population statistics has its limitations, which include the following:
- The data most likely does not perfectly represent the entire population since only people who own and use a mobile phone are included in the data (e.g. young children and elderly people will not be as well-represented in the data). Therefore, there is a discrepancy between the target population and mobile phone users - MPD captures only the subscribers, whilst the target population includes all of the individuals who reside in the country.
- The geographic, demographic, and socio-economic distribution of each MNO is likely to be different - e.g. an MNO with more expensive mobile plans is likely to overrepresent upper-middle income subscribers.
- Presence of IoT (internet of things) devices in the data set. Non-human generated data should be removed during data pre-processing by the MNO or algorithmically during the statistical production in order to avoid over-coverage issues.
- The use of several mobile devices and/or SIM (subscriber identity module) cards with either the same or different MNOs can also create over-coverage issues.
- Locational accuracy is determined by the density of the MNO's network - accuracy is generally higher in urban areas and lower in rural areas. The cell tower level is often seen as the most precise level for analysis, unless probabilistic methods are used to increase precision.
- A phenomenon named oscillation or "tossing" might indicate the movement of a subscriber between cell towers when in reality such a movement did not occur. Algorithmical solutions should be used to eliminate such occurrences of pseudo-movements.
- There may also be disparities due to the gender gap in mobile phone ownership, especially in low- and middle-income countries (GSMA, 2020b), and disparities between different socio-economic groups.
- More qualitative information such as socio-demographics might not be available or might not be of high enough quality in order to be used for official statistics.
Considering the aforementioned points, MPD should be viewed as a complement to traditional data sources, taking into account its own strengths and limitations.
