Social Media,Mobile Phones and Migration in Africa: A Review of the Evidence

I. Defining Methodological Boundaries: Evaluating the Evidence

The starting point to evaluate the evidence was guidance on ‘Assessing the Strength of Evidence’ issued by the UK Department for International Development (DFID) (2014). In recent years DFID, along with other government aid agencies, has pushed for the increasing use of systematic reviews and evidence assessments as both evidence to better inform their policies and programmes as well as indications of the efficiency and impact of these. More broadly, evidence reviews allow decision-makers and practitioners to act in an informed way and may provide academics with a clearer overview of the field to better interrogate gaps and possible areas of controversy (O’Leary et al., 2016: 75).

We adopted a semi-systematic approach ⁵ that combines elements of a systematic review, such as database searches and quality assessments, along with other methods such as snowball citation searches and expert recommendations. As the DFID guidance places greater emphasis on the evaluation of quantitative evidence (which flows from their focus on development economics), we sought inspiration from reviews that sought to engage with a diverse range of texts, much of which is qualitative (as the majority of research on social media and migration is), when evaluating the quality of evidence (Cummings et al., 2015; Schoemaker and Stremlau, 2014). Given the relatively low numbers of texts identified in our search, we strove to include as many pieces of valid research as possible. This necessitated the inclusion of grey literature, which are often hosted and disseminated in ways that are very different to academic publications and therefore may not be picked up by the database searches. ⁶ This is another reason why reaching out to experts, and especially practitioners working in the field, was central to our approach. In recognition of the challenges of including research by African researchers in semi-structured literature searches, and efforts to ‘decolonize’ review methodologies (Chambers et al., 2018), we explicitly sought out ways of incorporating African-led research into our review. We did this by reviewing university databases (where available) from some priority countries included in our search (including Ethiopia, Kenya, Libya, Nigeria, Senegal, Eritrea, Somalia) and by reaching out to researchers based at African universities.

In the first instance, relevant research was sourced through the systematic review process of undertaking a series of searches using variations of the terms ‘technology’, ‘migration’, ‘ICT’, ‘Africa’, ‘social media’, ‘mobile phones’ and ‘refugee’. These searches were carried out across multiple scholarly databases including Google Scholar, Web of Science and Scopus. The bibliographic data for all searches were combined into a single spreadsheet. Our semi-structured evidence review focusses on research published in the last decade between mid-2010 and mid-2020. The choice of this period is due to the fact that it spans both the period before and after the European ‘migration crisis’, as well as seeing a considerable expansion in both the use of and scholarship on social media. This period also experienced a significant increase in the use of mobile telephones across Africa. ⁷

In total, close to 26,000 texts were obtained through these database searches. From these, 4,074 duplicates were removed. The remaining 21,865 were then manually sorted based on the relevance of their titles and abstracts and of these 21,445 were later removed, leaving 420 publications. Texts were removed based on their lack of relevance, usually because they did not deal with the issue of human migration (removed examples included research on animal or data migration) or they focussed on migrants from regions other than Africa. These remaining texts then had their full abstracts considered, and where the abstracts did not provide an adequate amount of information on the content of the work, a reading of other sections of the text was then undertaken.

The use of empirical evidence was a requirement for inclusion and was the most restrictive factor eliminating many articles. At the end, 34 texts were relevant to this review following this initial surface-level analysis. ⁸ Five additional pieces of grey literature were added, out of the nine pieces in total which were reviewed. Despite their relatively small number, grey literature proved important as this is a rapidly emerging field and peer-reviewed academic publications often have a significant lag time between research and the publication of findings. Furthermore, given the strong interest from the policy community, including refugee support agencies as well as national security and international governmental bodies, these organizations have often been commissioning or initiating research to support their interventions. Grey literature, in this field, also tends to be more inclusive; we noticed a distinct trend that such publications are more likely to be authored or co-authored by researchers from African countries. While the publications may be sponsored or published by organizations in the global north, they often had partners from Africa. This research did not, however, differ from the positions taken by the entity commissioning the research. This is in line with previous studies have found that in some cases research on ICTs and development does not differ significantly depending on the geographical region where it is produced, but rather there is often a shared normative approach reflecting both the international policy agenda and funding realities (Gagliardone et al., 2015; Stremlau, 2018).

There are limitations to this method. Most notably is the choice of search terms, which we aimed to keep quite general to capture as many texts as possible initially, but we cannot rule out the possibility that there is other literature that uses different terms. Furthermore, we only used these search terms in English, ⁹ limiting the inclusion of literature in other languages. To triangulate these findings and to determine if there was any key literature that the systematic search may have missed, we contacted close to 70 experts (including both academics and practitioners) in the field(s) of ICTs and migration and asked them for their recommendations. We made considerable effort in targeting researchers from African countries. Experts were invited to provide up to five recommendations that focussed on the evidence they have relied on to understand the impact of social media and ICTs on the pathways of African migrants. We received recommendations of just over 50 texts, but there was less overlap than we expected between the texts that were collected via the systematic database searches and the recommendations of the experts.

As with all evidence reviews, one of the most challenging aspects is establishing a rigorous quality assessment protocol that to evaluate the relevance and rigour of the texts in relation to the question at hand. One consideration was scope, and most of the texts suggested by the experts did not directly engage with migrants of African origin. As mentioned before, and as identified in other more general scoping reviews on related topics (such as Mancini et al., 2019), much of this work relates to Syrian refugees, or mixed groups of migrants. A similar problem also arose in the consideration of the texts which were obtained via the database searches. While they made mention of migrants of African origin, very few offered empirical data that were directly relevant.

Only four texts overlapped between the 34 texts that were considered in full and the 51 recommendations by the experts, including: ‘The smartphone as a lifeline: an exploration of refugees’ use of mobile communication technologies during their flight’ (Alencar et al., 2019), ‘Social Media Usage, Tahriib (Migration) and Settlement among Somali Refugees in France’ (Charmarkeh, 2013), ‘A study of the communication channels used by migrants and asylum seekers in Italy, with a particular focus on online and social media’ (Sanchez et al., 2018) and ‘Becoming (Im)Perceptible: Forced Migrants and Virtual Practice’ (Witteborn, 2015). Of these texts, the first did not include African migrants in its sample and was therefore eventually excluded during the quality assessment phase.

There were several texts that were frequently recommended by experts. The most commonly cited publications, each with at least five separate experts recommending them, included: ‘Mapping refugee media journeys’ (Gillespie et al., 2016), ‘Smart Refugees: How Syrian Asylum Migrants Use Social Media Information in Migration Decision-Making’ (Dekker et al., 2018) and ‘Smart (phone) travelling: understanding the use and impact of mobile technology on irregular migration journeys’ (Zijlstra and Liempt, 2017). None of these was picked up by the systematic review process, as they were removed early on since they did not directly investigate populations that were African in origin. As with much of the literature that was recommended, these studies reflected more on the general research around migration pathways, ICTs and social media and tended to draw on examples from the Middle East where there is clearly a far larger corpus of research on this topic.

Overall, the evidence that emerged through the database search is sparse and points to some major systemic gaps. Most the papers that passed through the quality assessment filter relied on a combination of research methods, largely qualitative. Twenty-three papers used interview-based data with semi-structured interviews being the most common, while 11 papers were based on ethnographic observation, five used quantitative data and three involved focus groups. Other approaches included a range of digital methods, social network analysis, social media analysis, content maps and narrative techniques. Overall, many papers lacked methodological clarity and a significant portion of the papers offered a weak explanation as to the scope of data collected and how it was analysed.

In terms of the populations that were considered, the majority looked at more than one national group. There were some exceptions to this, with five texts considering Eritrean migrants only, three considering Somali migrants and one considering those from Senegal. Overall, migrants from the following countries were considered by at least one paper: Algeria, Burkina Faso, Cameroon, the Democratic Republic of Congo, Ethiopia, Gambia, Ghana, Guinea, Ivory Coast, Libya, Mali, Morocco, Niger, Nigeria and Uganda. Clear findings along national lines are not possible, because in many papers respondents were grouped together. Often these sample sizes were small (in some cases there was only one respondent from a particular country) or in some studies, the data concerning African migrants had been rolled into a larger data sample including migrants from other parts of the world, making these findings less useful for our purposes.

The debates and evidence that dominated the search can be grouped under three main themes: How social media encouraged (or discouraged) Africans to migrate to Europe (including its use in the pre-migration context); the role of social media during a migration journey; and evidence of ICT and social media use among recent arrivals in a destination (whether final or anticipated to be transient).

II. Exploring the Evidence

Many of the articles situated themselves in some of the broader debates that have historically characterized research on ICTs in Africa, including a focus on the ‘digital divide’ or the ever-fluctuating gap between information haves and have-nots. Over time, different characterizations of this phenomenon have emerged. In the early 1990s, the emphasis was on access to technology, or the lack of computers that are connected to the internet. This later evolved to focus on the divergent ways users were actually making use of ICTs, for example, how gender, economic background or age might shape internet use and the ability or reduced ability for someone to perform a task online (whether selling goods or looking for resources about migratory pathways or how to claim asylum). And more recently, studies on the digital divide have been connecting internet use with offline outcomes; for example, the relationship between how people may use the internet to search for information about migration and whether this leads to safer or more efficient journeys (Gagliardone, 2019: 63–79). One characteristic of this work that is helpful to keep in mind in the context of literature around new technologies and migration is the tendency for these technologies to been seen as disruptors, viewed in isolation, rather than embedded in a context where it may propagate continuities, or enhance certain social phenomenon but not entirely displace them (e.g. cutting the costs of procuring ransom incurred for criminal networks that are involved in holding migrants’ hostage, thus facilitating their business, but not necessarily empowering migrants to go around or cut-out these intermediaries).

The emphasis on ‘new’ technology has also advanced research approaches on migration that has been focussed on the ‘digital divide’ challenges migrants may have had to overcome prior to their journey, including limited mobile phone reception, the high costs of accessing data or the difficulty obtaining suitable devices. Differences in technological capabilities between migrants can exist across age groups, educational backgrounds and from both urban and rural regions, which can all affect the digital competencies of migrants prior to their leaving. One study explained how in 2017 only 1% of Eritreans had access to the internet (Van Reisen et al., 2019). National and community-level differences were emphasized by the Mixed Migration Centre’s study, which noted the significant discrepancies whereby:

migrants and refugees from Cameroon (78 percent), Nigeria (68 percent) and Togo (62 percent) have the highest rates of smartphone usage whereas those from Niger (65 percent), Burkina Faso (45 percent) and Senegal (41 percent) most often report using a non-smartphone. Not having access to any type of phone is most often reported by people from Burkina Faso (15 percent), Côte d’Ivoire (13 percent) and Senegal (11 percent). (Mixed Migration Centre, 2019)

In Haywood’s study, it was only the migrant originating from Eritrea who indicated that he had not needed a phone for navigation and communication, while all the remaining migrants who originated from regions of the Middle East indicated that they had used technology for their information needs (Haywood, 2016: 78). While the sample size of six is very small, it does reflect the absence of research looking at ICTs and social media in migration among African migrants and may also be indicative that for many migrants, longstanding communication and information networks have not necessarily been fundamentally altered by new technologies.

III. Conclusion

This article has sought to get behind the headlines—both positive and negative—about how ICTs are transforming migration. A focus on assessing and understanding the strength of evidence is an important aspect of more informed policy-making and more targeted interventions. In the context of humanitarian crises and migration, in particular, there is a tendency to jump on the latest technology and romanticize how it might transform the capacity of those intervening or of those migrating. This approach is also evident in the broader divides within the humanitarian tech sector, both those that have focussed on the individually empowering aspects of technology and those that have stressed how such technologies might be used to control and erode the privacy rights and autonomy of an already vulnerable group. There is a relatively robust and longstanding body of research around how African diaspora communities are using social media to engage in politics in their countries of origin (Gagliardone and Stremlau, 2011), and increasingly, the evidence is also suggesting that significant online hate speech, including in posts as well as Facebook Live or YouTube videos, is produced by individuals abroad adding a transnational dimension to addressing this content (Gagliardone et al., 2016).

Many of the seminal papers on ICTs and migration were picked up by both the systematic database search and the recommendations of experts. However, as mentioned previously, much of this work has focussed on refugees from Syria and Afghanistan, and so was not directly relevant to the scope of this review. These works offer a point of comparison, especially since they have been generalized as representative of the experience of other migrant communities. Many used Syrian refugees as their case study showing how they use social media information both before and during their journeys (Dekker et al., 2018: 9), as well as illustrating how these technologies being empowered migrants to better plan their own journeys (Zijlstra and Liempt, 2017: 183). They elaborated how mobile phones are ‘both a resource and a threat’ (Gillespie et al., 2016: 2) and focussed on some issues associated with the arrival in Europe, including how ‘the appearance of digitally connected refugees was perceived as incongruent with Eurocentric ideas of sad and poor refugees fleeing from war and atrocities’ (Leurs and Ponzanesi, 2018: 6). Overall, this literature emphasized how these technologies can fulfil a range of purposes, including ‘as a companion, an organizational hub, a lifeline and a diversion’ (Alencar et al., 2019: 841). As this systematic review has illustrated, however, there is significantly less evidence available on the use of ICT and social media in African migration pathways. This semi-systematic review has been unable to find clear evidence of a general reliance on ICTs and social media during all stages of a journey. This is not to say that it does not exist, but it highlights the considerable gaps within the literature and demonstrates how ungrounded many assumptions may be, particularly when they involve interventions such as social media campaigns aimed at discouraging migrants to leave the continent or strategies to provide for ‘information needs’ for recent arrivals.

Finally, one unintentional but significant finding of our evidence review has been the extent to which the literature that formed our final corpus for analysis came from western scholars. We explicitly tried to address this in the methodology by reaching out to experts and centres on the continent, and search through depositories of African theses on the continent, but research production on migration (particularly Africa–EU migration) remains centred on the global north. While a number of the African migration centres have examined the role of media (and sometimes social media) they have tended to focus on local or national case studies. ¹² This reinforces calls for diversifying, decolonizing and strengthening the opportunities for African research centres and universities to take a leading role in empirical research production across the continent, and in Europe, as well.