What is 4Mi Migration
Data? 

4Mi started in East Africa (Nairobi) in 2014 and was conceived by the Regional Mixed Migration Secretariat (RMMS, which is now the East Africa & Yemen hub of the Mixed Migration Centre), to fill information gaps around mixed migration, especially protection risks and vulnerabilities of people on the move.

4Mi is MMC’s flagship primary data collection project. It is run by MMC staff and 4Mi data feeds into a large number of MMC’s knowledge products. MMC is part of and is governed by DRC. MMC’s institutional link to DRC ensures its work, including that of 4Mi, is grounded in operational reality and can adapt to those realities. At the same time, MMC acts as an independent source of data, research, analysis, and policy development on mixed migration. The position of MMC does not necessarily reflect that of DRC. In different countries or along routes, MMC (often including 4Mi teams) works alongside DRC in joint programmes to provide evidence on the situation of people in mixed migration.

Yes, 4Mi work with partners in different ways. 4Mi works with partners to conduct data collection, where this approach is deemed more viable and effective than directly recruiting enumerators ourselves. These may be local associations, NGOs, private companies or social enterprises, and there is a thorough assessment process to evaluate the partnership before proceeding. 4Mi also develops more partnerships that go beyond data collection, designing projects and developing methodologies with both international and local experts and institutions. And finally, 4Mi often shares its data with partners or contributes its data and analysis to specific products and projects.

4Mi conducts continuous data collection with a broad and heterogeneous target population, and its research questions are broad. More precise questions are formulated when conducting specific analyses of the data. The broad underlying questions are:

• Who are undertaking mixed migration journeys? 
• What do refugees’ and migrants’ journeys look like: route, duration, conditions, financing? 
• What is motivating them to make the journey?
• What are their aspirations and intentions?
• What dangers do they face on the journey and where?
• How vulnerable are refugees and migrants on their journey, and what determines their vulnerability?
• How prevalent is smuggling and how does it manifest itself?
• How are decisions made regarding migration and the migration journey and what influences those decisions?

4Mi complements other data collection projects because it:

• collects original first-hand data directly among refugees and migrants; 
• collects in-depth yet quantitative data on the range of refugees’ and migrants’ experiences;
• collects data continuously, ensuring up-to-date information and enabling a degree of comparison over time; and
• uses the same tool across all regions and routes, enabling global and cross-regional comparison.

4Mi data is collected directly from refugees and migrants, making it an invaluable source for understanding the details of their journeys, perceptions, and perspectives at scale. 4Mi data and analysis means that migration policies, debates, and protection responses for people on the move can be grounded in evidence from migrants themselves.

4Mi is continuous, and for the most part, interviews respondents once on their journey. Each respondent reports on their entire journey, providing a history. And 4Mi enumerators are often operating at different points along a mixed migration route, providing data on the dynamics at different points on the journey. However, none of this is strictly longitudinal data.

Beginning in 2023, after successfully developing a longitudinal approach to data collection in 2021, 4Mi will run a number of projects that follow a select number of subjects (respondents) over a longer period of time, interviewing them over regular intervals.

4Mi takes a quantitative approach to data collection, while at the same time providing rich and in-depth insights through the kinds of questions it asks. It uses standardised, structured surveys with almost exclusively closed-ended questions. This provides measurable data, which can be used in statistical analysis. Due to the type of data (much is categorical) and the sampling approach. MMC often restricts its analysis to descriptive statistics. However, it is possible to carry out more advanced statistical analysis with some 4Mi datasets.

In addition to the core migrant survey which covers the full migration journey,  special modules can be added to focus on specific themes or emerging issues (for example in 2025 a module was added looking at the measures by the Trump administration). These modules are integrated into the core survey for rapid and targeted data collection using the already existing 4Mi infrastructure to avoid the need to set up a new standalone research project. 

Other surveys have also been developed for specific thematic 4Mi projects. For example, during Covid-19, a new survey captured the pandemic’s impact, with over 25,000 remote interviews collected in the first year, and 4Mi Cities was developed to explore migration realities in urban hubs. 

On selected routes, 4Mi also implements a longitudinal approach, repeatedly interviewing the same respondents to track how journeys, decisions, and risks evolve over time.

Finally, 4Mi has developed targeted approaches to reach specific populations in addition to people on the move, such as developing surveys of smugglers and surveys of returnees.

4Mi utilizes a mix of purposive and snowball sampling methods, preferring direct purposive sampling. That is, potential survey respondents are selected at key locations according to a small set of criteria.  Selected respondents often refer others who also fit the criteria, and so on. Enumerators also identify respondents through referral by community leaders or people working with migrants, as well as inviting or approaching potential participants via social media platforms (although we do not use advertising).

All these methods are strictly managed to ensure that participation is voluntary, and that bias in sampling is mitigated to the extent possible.

Due to the sampling methods used, 4Mi data is not representative of national or international migration dynamics. It therefore cannot be used to provide estimates of the volume or characteristics of the overall migrant population. 

Instead, 4Mi seeks to maximise access to the population of interest, so that it can include the full diversity of that population. It does this through the quantity of interviews (more than 10,000 interviews are conducted each year), careful scoping and selection of sampling sites, and careful recruitment of enumerators. For example, 4Mi policy is to recruit a minimum of one male and one female enumerator in each location, as this ensures a better male: female ratio of respondents. 4Mi also implements a sampling schedule in some regions, to ensure an even presence of enumerators across recruitment locations and recruitment times at each location. Specific projects, with more specific populations of interest, allow for a more random approach, where we can be confident that the data is more representative. A note on sampling is included in all MMC’s analysis products.

The main 4Mi survey consists of an in-depth structured questionnaire of around 160 items. Most items are questions, and a range of measures (nominal, ordinal, and interval) and answer options (forced-choice, multiple-choice, and Likert scale) are used to record the respondents’ answers.

The core 4Mi survey questions cover 8 key areas:

1. Profile: age, nationality, religion, sex, socio-economic status, etc.
2. Route and journey conditions: country of departure, transit countries, number of people travelling together, etc.
3. Drivers and aspirations: reasons for leaving, decision-making, determination to reach destination, etc.
4. Protection risks: perceived risks and dangerous locations, perpetrators, etc.
5. Assistance: type of assistance needed, assistance providers, locations where assistance is most needed, etc.
6. Smugglers: services provided by smugglers, payment arrangements, migrants’ perceptions of smugglers, etc.
7. Financing the journey: sums paid, access to money while travelling, bribes paid, etc.
8. Access to information: sources of information used before and during the journey, phone and internet access, social media use, etc. 

Respondents are asked one open-ended question, to allow them to speak more in their own words about any particular aspect of their migration experience. 

Finally, the questionnaire includes questions (some for the enumerator) on selection criteria and other metadata covering interview conditions, possible sources of bias, etc. Respondents are also invited to ask questions at the end. 

The core 4Mi survey is the same across all regions, to enable comparison and ensure consistency. No questions can be removed from the core survey,  but optional modules cover specific themes in greater depth and are added to the survey in particular regions, for a particular time period, or for particular population groups. Modules have been developed on smuggling, children and youth, unexploded ordnance, livelihoods, future intentions and decision-making, social media, and exploitation.

Enumerators use the same protocol to interview respondents across all regions. First, enumerators read a script to respondents, informing them about DRC and MMC, the purpose and scope of the interview, and that the information shared by them is voluntary, anonymous and confidential. 

Enumerators are trained to ensure that interviews are conducted in places where both they and the respondent feel comfortable, often a public place that offers a degree of privacy. They strive to conduct the interview uninterrupted, but it can take place on more than one occasion. 

The enumerator uses a smartphone to administer the questionnaire, reading out the questions and (where prompting is required) answer options. At the end of the interview, enumerators record any challenges during the interview.

4Mi interviews are most often conducted face-to-face, but where this is not possible, most often for safety reasons, the interviews are conducted remotely, by phone. 

Sampling is through a mixture of purposive and snowball approaches, and participants are recruited directly as well as through a number of remote or third-party mechanisms, using social media, community networks, and assistance programs. The teams have developed mechanisms that are relevant to the local context to promote diversity in sampling and mitigate sampling bias.

Data protection measures are in place to ensure that the data collected remains anonymous. Enumerator training also reflects the different requirements of telephone interviewing.

The written questionnaire is available in Arabic, English, Dari, French, Pashto, Somali and Spanish. To reach a more diverse sample, where possible, some enumerators simultaneously translate the survey into the respondent’s first language while administering it.  

Our glossary comprises definitions of terminology used in the survey to ensure that translations are accurate and that participants, enumerators, and analysts all have a common understanding.

All enumerators and staff involved in 4Mi are trained in research ethics. 4Mi deals with potentially sensitive topics and enumerators will engage with participants who may have been exposed to physical and sexual abuse or other traumatic experiences. It is therefore vital that enumerators are trained properly on how to approach participants and that they understand the importance of informing participants about the purpose of the interview and listen, show empathy and respect the limits of what the respondent wishes to share. 

All participants are informed orally by the enumerators about the purpose of the data collection, the use of the data, the anonymity and confidentiality of the survey interview, and the right not to respond and to withdraw. When discussing particularly sensitive issues relating to protection, we focus on respondents’ perceptions and do not probe their personal experience. At the end of each interview, time is set aside for participants to ask questions, making sure that participants are not left with unaddressed concerns. Participants are given an email address to contact if they have any questions after the interview.

All interviewees are anonymous and once an enumerator has uploaded a questionnaire to the 4Mi server, 4Mi has set procedures for data management and protection, and will not use the data for any purposes beyond those for which the participants gave their consent. 

4Mi can also take an emotional toll on enumerators. Their training includes measures on self-care, and the 4Mi teams maintain regular follow-up with the teams so that any concerns are identified and addressed.

The enumerators who conduct the structured interviews are either from among the migrant community themselves, or (in some regions, depending on profiles and language skills) they can be nationals of the country in which the interviews take place. Attention is paid during recruitment to enumerators’ profiles (e.g. age, sex, languages spoken), as this contributes to the diversity of the sample.

This depends. In 2023, the enumerators were from 26 different countries. Some are migrants themselves, but approximately 63% of enumerators are nationals in the country where they operate.

The enumerators receive a specific training following recruitment. Where possible, this is face-to-face and run by MMC staff. As well as in-depth sessions on administering the survey, training covers the concepts around mixed migration and primary data collection sampling methods, and ethical considerations. MMC staff provide continuous quality control and feedback during data collection.

Most enumerators are under contract to DRC. This can be a service contract or an employment contract, and varies by location. Enumerators are most often paid per questionnaire or at a daily rate. Where DRC cannot support enumerators directly (most often because DRC is not registered in the country of operations), MMC works with partners, who contract enumerators.

4Mi’s target group for the core survey is adults on the move away from their country of departure on mixed migration routes, irrespective of status, though often engaging (at least for parts of the journey) in irregular migration. For particular projects conducted at regional or country level, the target be more specific. For the smuggler survey, 4Mi targets individuals who facilitate irregular movements of people. See our definitions for more details. 

For the standard migrant survey, 4Mi uses the following criteria to select respondents:

• Respondents must be 18 years or older
• Respondents must have crossed a border
• Respondents must have arrived in the location of interview no longer than 2 years ago (this can increase according to context and access to the population) 
• Respondents must not be returning to their home country or country of departure
• Respondents must not have been interviewed by 4Mi before

In some locations, or for some 4Mi projects, we may target particular nationalities or other profiles in addition to these criteria. In other 4Mi programmes (e.g. 4Mi Returns), the criteria differ.

4Mi is a data collection and research tool and does not provide assistance. 4Mi enumerators are not trained to provide assistance and MMC is not an operational organisation. However, while maintaining a necessary boundary between data collection and research on the one hand, and provision of assistance on the other, MMC has developed an orientation mechanism, providing information on where to find help. This is being implemented in some locations.

When conducting most data collection, respondents’ names are not recorded, and instead an ID number is automatically generated. 

When conducting remote data collection (by phone), personal data is collected, but is stored separately to survey data, and MMC has procedures for its management and protection. Enumerators cannot access survey data once they have submitted it to our data collection platform. 

When conducting longitudinal data collection, it is necessary to collect identifiers in order to be able to re-contact participants. The storage and processing of this data is subject to rigorous data protection protocols. The data is pseudonymised as soon as possible and anonymised as soon as data collection is complete.

As the 4Mi survey is anonymous, results cannot be shared directly with the respondents. That said, respondents are told about MMC and its website, where results are published, and an email address is shared in case they wish to contact us after the interview.

Yes. Respondents can ask questions and make comments at the end of the interview, and these are recorded. They are also given a contact email for later feedback.

4Mi data is not representative of the general migrant population (see question under methodology for more) and cannot be used to estimate the numbers of migrants in or on the move in the regions where it operates. 

4Mi data is also self-reported and MMC has no means to verify, for example, reported incidents.

4Mi data and analysis provide contextual insights and evidence to support new thinking about the focus and location of new humanitarian or development activities, new directions for migration and related policies, and increases the knowledge base on mixed migration generally.  

4Mi data is used in most MMC reports and publications, as well as in 4Mi data explorer and MMC frequently presents 4Mi findings at conferences and seminars. MMC has a number of agreements with partners to provide 4Mi data and analysis, and frequently responds to ad hoc requests for information, both from DRC staff and external organizations. 4Mi data has been used in a wide range of non-MMC projects and reports, including IOM’s “Missing Migrants” project and in reports by organisations such as UNICEF, UNHCR, UNFPA, UNODC, OHCHR, Clingendael and the Expert Council of German Foundations on Integration and Migration, as well as governments.

Yes. 4Mi can provide the necessary contextual analysis, and provide credible evidence on where and why there is a need to develop a response for migrants on the move. It can provide the project rationale, and the baseline and context it provides can inform the development of more targeted needs assessment, independent of 4Mi. In terms of the programme cycle, 4Mi plays a particular role in the initial analysis, strategic design and (as a consequence) resource mobilisation. Given the limitations of the methodology, 4Mi cannot provide rapid data on the need for real-time flexible adjustments during programme implementation, though during longer-term programmes it can provide indications of shifting trends and needs.

4Mi data and analysis can inform humanitarian response programming, policy making, and additional research into mixed migration. Use 4Mi data explorer to explore the data and generate findings that are useful to you. Graphs and data published in MMC publications or generated through 4Mi data explorer can be used freely in external publications or presentations, as long as the source is acknowledged as “4Mi data – Mixed Migration Centre”. MMC also has an information request system: contact us for details.

MMC has a number of 4Mi data sharing agreements with selected organisations and responds to ad hoc requests for parts of the dataset and/or data analysis on a case-by-case basis and depending on the nature of the request and available resources. MMC does not publish or share the full raw data set.

After an enumerator submits a completed interview to the 4Mi online system, the raw data is stored on servers in Germany. Data that is exported is stored locally, in one of the MMC hubs, and periodically duplicated and transferred to servers housed in Europe.

Data is then subject to quality control. Survey validation procedures minimize the risk that fraudulent or fabricated data, or even errors in inputting, are introduced in the 4Mi dataset. For example, interviews that were completed in less than 20  minutes or that do not match the interview locations assigned to the enumerators are discarded. 

Data cleaning involves verification of internal inconsistencies in responses and unclear responses. 

Data analysis varies according to the size of the dataset and the task. 4Mi snapshots are produced by regions, and generally present descriptive statistics as visuals. More in-depth studies use regional or cross-regional datasets and can include inferential statistics (regression analysis, etc.).