What should we call the growing presence of data in international development?
That’s a question I posed on the ICT4D Facebook group.
Though #datarev is a popular hashtag, “data revolution …” did not arise, and just as well – it is naive hyperbole to suggest data is going to transform development structures.
The proposed terms fall into four orientation categories.
1. Goal-oriented terms. The main one here is “data for development” which is admirable in focusing on the purpose of the data, and in offering a ready-made acronym – D4D – which I’ve talked about earlier. It’s moderately-popular, partly thanks to Orange’s D4D Challenge, and has a nice continuity with ICT4D. The term is new, but the main problem is its failure to reflect the changing role of data in development – data has always been used for development purposes.
2. Facilitation-oriented terms, especially “data-enabled development” (DED) (data-facilitated, data-catalysed as synonyms). This has the same problem as D4D: per se, the term gives no sense of the change that has occurred. And DED has no presence in the field as a term.
3. Impetus-oriented terms, especially “data-driven development” (DDD) (data-centric as a synonym). This has some presence in the field, though less so than D4D, with – for example – a World Economic Forum group and report on DDD, and the Global Partnership for Sustainable Development Data having some commitment to the term. I’m guessing this will become more widely-adopted – “data-driven” already has Wikipedia entries for equivalents such as data-driven journalism. However, it rings many alarm bells in placing too much deterministic emphasis on data as an agent in development – put simply, people not data drive development.
4. Change-oriented terms, especially “data-intensive development” (DID) (data-rich as a synonym). The great thing about this term is that it explains what is new and different – that data is playing a greater role in development decisions and processes – without so-much falling into traps of determinism and value judgement. I think “data-intensive development” is the most appropriate of the terms on offer. As yet it is little-used, so the only way is up . . .
If you’ve got a better suggestion, you’re welcome to say what it is and why it’s better.Follow @CDIManchester
What can actor-network theory offer to our understanding of technology and development?
This blog entry summarises the answer from an open access paper in the journal Development Studies Research: “Technological Change in Developing Countries: Opening the Black Box of Process Using Actor–Network Theory”, and it builds on an earlier entry on ANT and development.
Technology rather dropped from the development agenda during the 1980s and 1990s, but has re-emerged strongly in the 21st century; not least due to the spectacular diffusion of ICTs.
Yet, to date, conceptualisation of technological change in developing countries has had three problematic gaps:
- It has been de-humanised: organisations are recognised as actors but people – as identifiable individuals with agency – rarely appear in the technology and development literature.
- Technology may be understood as a physical artefact, as a system of elements, as the embodiment of knowledge. But it is not seen as playing any active role: technology is acted-upon but is not itself acting.
- Research has tended to study factors or social structures affecting processes of technological change. But it does not describe those processes in detail: actual practices of change tend to be black-boxed.
In sum, research to date has typically stood outside the technology processes it seeks to investigate; freezing them in time and concealing their main actors.
As luck would have it, these are just the kind of lacunae that actor-network theory was intended to address. Yet application of ANT to cases of technological change in developing countries has been rare; and within development studies literature, almost non-existent. So new ANT-based case studies of technology and development are required to assess what insights actor-network theory can offer.
One such case study – applying Callon’s “moments of translation” to a digital information system in the Sri Lankan public sector – is presented in the Development Studies Research paper (which should be accessed for full details). It finds that an initial network supporting technological change fell apart in mid-project, and had to be reconstructed around a new technology design and a new vision for future change.
Three challenges emerged in applying ANT:
- Methodological: admission of subjectivity in framing an ANT-based case, and problems of thinning out detail to fit a journal-length account.
- Analytical: that ANT can provide a rich description of how things happen, but stutters in seeking to analyse why.
- Instrumental: the difficulty of extracting practical guidance from ANT other than rather “Machiavellian” prescriptions.
On the other hand, the case analysis shows that ANT can open the black box of technological change processes and offer new insights:
- Networks: explaining the networks of relations that both support and oppose technological change, and also the detailed process by which they come to be formed, dissolved, etc.
- Technology: exposing the active role that technology plays in international development – shaping, enabling, co-operating, resisting, etc.
- Human practices: providing a detailed account of the role played by individuals and groups in technological change; particularly the way in which lead actors modify the perceived interests and even identities of others involved.
ANT therefore shows us not just that human interests, identities and relations change in a technology-and-development project; it also explains in what way they change, how it is that those changes come about, and how they relate to the project’s trajectory.
The case analysis shows that ANT will not help answer questions about the impact of context on technological process, or about the developmental impact (in the traditional sense) of technology. However, it may help to answer questions such as:
- How do we explain the trajectory of a technology and development project?
- How does a particular innovation in a developing country diffuse, scale up or sink without trace?
- What role does technology play in processes of technological change?
- How does power manifest itself in such processes? How are apparently relatively powerless actors sometimes able to influence the direction of technological change? How are apparently relatively powerful actors sometimes not able to get their way on a technology project?
As the technology used in development becomes more complex, more interconnected, more intertwined into the lives and livelihoods of developing communities, and changing at an ever-faster pace; then ANT will likely become more relevant and more useful as a conceptual frame.Follow @CDIManchester
The spreadsheet linked below provides time-series data for India’s IT industry, updating data from an earlier blog entry on Indian IT data to 2009. Software export figures run from 1980; overall IT outputs from 1991; and detailed breakdown from 1998 including BPO (business process outsourcing) data from 2000. Data from 2009/10 uses a different source, so changes from 2008/09 to 2009/10 are not reliable.
a) Indian Software Exports
a1) Indian software exports are huge – roughly US$75bn in 2014/15 (and c.US$100bn if BPO services are included) – and continuously registering double digit annual growth.
a2) But the overall pattern of growth is slowing: the ten-year annual growth average was c.40% in 2002; c.30% in 2008; c.20% in 2014.
a3) IT software/services’ share of total exports remains roughly static: it was just under 14% in 2003/04 and just under 15% in 2013/14.
a4) Market diversification for Indian software remains limited. In the early 1990s, export destinations were: US (c.60-65%), UK (c.10%), other Europe (c.10%), Aus/NZ (c.5-10%), Asia (c.0-3%). Twenty years later in 2013-14, export destinations were: North America (63%), UK (13%), other Europe (11%), Aus/NZ (4%), Asia (6%).
a5) Location of production has changed. In the early 1990s, 75% of work took place on-site, 25% in India. By 2013/14, it was said that 20% of work took place on-site, 80% in India. This means that net foreign exchange earnings will have risen as a proportion of gross since offshore work requires much less foreign exchange outflow than on-site working.
a6) One source claims that productivity (as measured by average revenue per employee) in the Indian software sector has risen from c.US$7,000 per head in the mid-1990s, to c.US$16,000 in the late 1990s, to US$38,000 in 2014. But my own data gives a completely different picture: that productivity in the 1990s was static at just over US$30,000 per head, and thus has risen very little during the 2000s: at best by 1-2 percentage points per year.
b) Domestic IT Production
b1) Although the Indian domestic IT market is large and growing, production for exports is growing faster than production for the domestic market. As a result, the share of exports in total IT output has risen from 19% in 1991/92 to 49% in 2000/01 to 67% in 2007/08 to 81% in 2014/15.
b2) IT production for the Indian domestic market and domestic IT consumption are very different. For example, domestic computer hardware production in 2013/14 was roughly US$3bn. But domestic IT consumption was US$12.4bn. In part, this may be because the two figures are counting different things (e.g. consumption figure includes peripherals, network kit, storage, etc). But it likely also points to a high – and said to be growing – share of imports in Indian domestic IT consumption.
c) IT Sector Overall
c1) The IT sector overall in India represents just over 5% of GDP in 2014/15.Follow @CDIManchester
 Mani, S. (2014) Emergence of India as the world leader in computer and information services, Economic & Political Weekly, XLIX(49), 51-61
 Heeks, R. (1996) India’s Software Industry, Sage, New Delhi
 Heeks (ibid.)
 RBI (2015) Survey on Computer Software & Information Technology Enabled Services Exports: 2013-14, Reserve Bank of India, New Delhi https://rbi.org.in/scripts/BS_ViewBulletin.aspx?Id=15452
 Malik, A. & Nilakant, V. (2015) Context and evolution of the Indian IT industry, in: Business Models and People Management in the Indian IT Industry, A. Malik & C. Rowley (eds), Routledge, Abingdon, UK, 15-34
 Heeks (ibid.)
 Chawla, M. (2014) Indian IT hardware markets stands at $12.43bn, The Economic Times, 25 Jun http://articles.economictimes.indiatimes.com/2014-06-25/news/50856134_1_anwar-shirpurwala-biswapriya-bhattacharjee-indian-it
e-Government develops over time. Researchers want to track this; practitioners want to benchmark where they are in relation to others. The result has been the development of e-government maturity models. But there are difficulties with the most popular models, and here I propose an alternative.
There are many e-government maturity models: Fath-Allah et al (2014) provide a helpful overview that identifies and then analyses 25 different models. By far the best-known is Layne & Lee’s (2001) four-stage model (see below): not just the most highly-cited model but the most highly-cited of all e-government papers.
This foundational model has been revised (e.g. Andersen & Henriksen 2006) and revisited (e.g. Lee 2010). Drawing on this past literature but also additional insights, three key challenges to the Layne & Lee model can be identified:
1. US-centricity. All models are a product of their context. Layne and Lee’s model derives solely from e-government experience in the US. As a result – even 15 years after its development – the model focuses mainly on higher levels than actually achieved by most governments worldwide; at least judged on the basis of the UN e-Government Survey (e.g. UNDESA 2014).
To address this, it will be useful for a revised maturity model to disaggregate two pre-transactional stages; one related to one-way publication of government information; one related to two-way interactive capabilities of e-government.
2. Beyond Integration. All models are a product of their time. At the time Layne and Lee were writing, it was perfectly reasonable to identify integration – joining up e-government services vertically across levels of government, and horizontally between sectors of government – as the pinnacle of e-government’s achievement: the completion of all of a citizen transaction online. Subsequently, however, e-government has moved beyond this. Maturity model designs that move beyond integration are therefore the focus for a number of the revisions charted by Fath-Allah et al.
One example now visible is government undertaking a transaction proactively; e.g. proactively reissuing a passport when the old one nears renewal or an auto license with payment set up via an automatic direct debit mandate. Or, governments now customise processes so that services are individualised; e.g. providing personalised information matched to the known needs of the individual citizen. Or, governments may even eliminate a process, no longer requiring anyone to do it; e.g. government proactively issuing a birth certificate from hospital records so the citizen no longer needs to register a birth.
These types of further development need to be incorporated into any revised maturity model.
3. Disaggregated Dimensions. All models are a product of their mindset. Not just the Layne and Lee model but also its suggested variants have a single-path perspective on e-government maturity. At first glance this can be hard to notice since the Layne & Lee model plus some variants is illustrated within a two-dimensional space. But that space is not used: these models are functionally-equivalent to a one-dimensional ladder model.
As a result, there is no differentiation between the “front-office” nature of the interface / interaction, and the “back-office” nature of the underlying process. Yet these can progress at different rates. Some e-government applications are quite transactionally complex but not integrated with other services. Other e-government applications – such as a number of government portals – are simple in interactional terms but integrate across all of government.
A revised maturity model should take this into account, and provide a truly two-dimensional space within which e-government can mature.
The Manchester e-Government Maturity Model
The Manchester e-Government Maturity Model seeks to incorporate the responses to these three challenges:
- It reflects current global realities of e-government by having separate categories for “Informing” (one-way publication of information) and “Interacting” (two-way communication between government and users).
- The extent of change in underlying government processes includes the Layne and Lee components via categories of “Digitisation” (simple automation of existing processes) and “Improvement” (bringing processes together via horizontal or vertical integration). But it then moves beyond these to more fundamental process change via categories of “Redesign” (such as proactive or customised transactions) and “Transformation” (complete reworking of processes such as their elimination or reversing from government- to citizen-led).
- Delinking front-office interaction from back-office process change.
The figure below shows the model, including some examples at various different points in the maturity space; with growing maturity understood to be movement from bottom to top, and left to right.Follow @CDIManchester
 Fath-Allah, A., Cheikhi, L., Al-Qutaish, R.E. & Idri, A. (2014) e-Government maturity models: a comparative study, International Journal of Software Engineering & Applications, 5(3), 71-91 http://airccse.org/journal/ijsea/papers/5314ijsea06.pdf
 Layne, K. & Lee, J. (2001) Developing fully functional e-government: a four stage model, Government Information Quarterly, 18(2), 122-136
 Andersen, K.V. & Henriksen, H.Z. (2006) e-Government maturity models: extension of the Layne and Lee model, Government Information Quarterly, 23(2), 236-248
 Lee, J. (2010) 10 year retrospect on stage models of e-government: a qualitative meta-synthesis, Government Information Quarterly, 27(3), 220-230
 UNDESA (2014) United Nations e-Government Survey 2014, UN Department of Economic and Social Affairs, New York, NY http://unpan3.un.org/egovkb/Portals/egovkb/Documents/un/2014-Survey/E-Gov_Complete_Survey-2014.pdf
Openness and transparency are good things and the more we have of them the better. Right? Wrong.
In contexts of too little openness – “hypo-transparency” – ICTs can help bring greater transparency, with positive developmental effects. But in contexts of relative openness, ICTs are ushering in a hyper-transparency that will destroy public institutions. As summarised in the figure below, I therefore propose an inverse-U relation between e-transparency and various measures of political development, such as trust in public institutions.
As an experiment, try the following. View your beloved from a very far distance. They are a tiny speck, and you feel nothing for them. Now move closer to view them from a few feet away. Likely you will see much to admire and feel a warm glow (if not, it may be time for an upgrade). Now get up really, really close and examine them in minute detail – take a look up their nose, in their ears, inside their . . . well, you get the idea. That glow’s probably not quite so warm now, is it?
Something similar happens with ICTs and transparency. Applied in corrupt, opaque, self-serving environments, ICTs have been shown to reduce corruption and improve the efficiency and equity of practice. But applied further in democratic environments where a reasonable degree of e-transparency and openness already exists, ICTs can make things worse rather than better.
Through greater e-transparency, ICTs help us know ever-more about the behaviour (decisions and actions) of those within public institutions. The majority of that behaviour will be appropriate. But humans are flawed, so they will always make mistakes, act selfishly, and do bad things. Absent other effects, the greater the transparency, the greater the absolute amount of such inappropriate behaviour that will be revealed, and the less citizens will value and trust public institutions.
Any effects of transparency in reducing the amount of behaviour that is inappropriate are mitigated both internally and externally. Internally, transparency pushes institutions to spend increasing time on non-value-adding defensive activities. These include trying to second-guess and avoid what might cause offence or other negative public reaction; excessive caution in behaviour to avoid risk or failure; and inefficiencies in protecting necessary confidential interactions – the “safe space for genuine deliberation” – from external gaze. Yet, “without the exchange of confidences, it is not possible for people to have real confidence in their colleagues and in the organisations that employ them”.
Externally, ever-greater flows of e-transparency data undermine public institutions because of . . .
- Cognitive deficits: the greater the flow of data, the lower the absolute availability of knowledge and motivation among the public to properly interpret that data, leading to a dominance of simplistic interpretations, many of which are negative because of . . .
- Cognitive bias: the negativity bias that causes humans to attend more to bad than good news, to remember bad more than good news, and to form negative stereotypes more quickly which are more resistant to disconfirmation. And the tendency, for example when searching online, to attend to extreme rather than average data. Extreme and negative interpretations of data on public institutions become more prevalent because of . . .
- Political incentives: attention and profile online accrue to those who posit more extreme views, and there are plenty of commentators who have political or economic incentives to criticise current public institutions and who – within already-relatively-open contexts – are able to do so. They have an ability to shape the narrative in part because citizens give up their own interpretation due to cognitive deficits. And thus we have a self-reinforcing spiral.
The impact of this can be seen, for example, in the decline of trust in public institutions in democracies during the Internet era. Dating this from the turn of the century, some illustrations:
- A decline in trust in the European Union from 40% to 30%.
- A decline in Americans’ confidence in Congress from 27% to 7%.
- A halving of Americans’ confidence in news media.
Of course e-transparency is not the only factor behind trust, but a review of some key literature finds little evidence that transparency builds trust. Instead, “in a number of cases, the evidence points in another direction: that is, transparency may ultimately decrease trust”.
This has a number of negative knock-on consequences if lack of trust leads to calls for greater transparency which leads to a further erosion of trust. With only a minority – sometimes a small minority – of citizens trusting institutions, those institutions are weakened in their ability to defend the public realm and public interests. And we see a shift in power from public to private institutions, and from centrist to more extreme political views and parties.
Is this an argument against e-transparency? It is not. But it is an argument that:
- We are guided by the inverse-U curve to give highest priority to using ICTs to open up the most-powerful, least-transparent institutions. That means authoritarian regimes and transnational corporations. Oh, and FIFA. Don’t applaud Edward Snowden until he exposes the workings of the 3PLA, or Julian Assange until he leaks the tax avoidance plans of global IT firms. If you want a transparency hero, pick Herve Falciani.
- We place greater emphasis on accountability than transparency. Transparency, in Furedi’s words, fosters “a political culture of voyeurism”. Accountability – at least when properly designed – fosters reasoned, considered checks and balances against abuse of power.
- We accept there are limits to openness, and that we want transparency but not hyper-transparency: “A democratic society should understand that it is important to uphold the right to the private exchange of views and that not everything officials do ought to be visible to all”.
 Furedi, F. (2011) Let’s stop kowtowing to the cult of transparency, Spiked, 5 Oct http://www.spiked-online.com/newsite/article/11140
 Furedi ibid.
The UN Secretary General’s Synthesis Report on the Post-2015 Agenda was released on 4th December. It’s just one document but could be bellwether of future development priorities.
It represents the culmination of a historical trajectory in the relative presence of “ICT” vs “data” in the development discourse. As discussed in a more detailed post-2015 vs. MDG agenda analysis, ICTs outpolled data at the turn of the century in the Millennium Development Goals. In early post-2015 development agenda documents, this reversed – data was mentioned three times more than ICTs. In the Synthesis Report, the ratio is close to 10:1. Data is mentioned 39 times; ICT just four times.
What would it mean if data replaces ICTs as the core focus for informatics in international development?
For many years there have been concerns about the techno-centricity of ICT4D: the assumption that technology, alone, can be sufficient to generate development; and the failure to recognise the wider contextual factors that govern the impacts of technology. Moving to a data-centric view helps a bit: it moves us to think about the stuff that technology handles, rather than the technology per se.
But it doesn’t help a lot. As Information Systems 101 teaches, it is information, not data, that has value and adds value. And a data-centric view is not inherently better than a techno-centric one at recognising the importance of context. For both these reasons, as I’ve discussed earlier in this blog, it looks like many “data-for-development (D4D)” initiatives to date are stuck at the very first upstream step of the process – they produce data but only rarely produce results.
For the academic community working in the sub-discipline of development informatics, a relative shift from ICT4D to D4D will mean a requirement for new research focus and skills. At the least, we will need to add new research projects and research competencies around data and decision sciences. At the most, these might partly replace – at least in relative weight – technical computing activities and capabilities.
That reorientation will certainly be true of the practitioner community, leading to demand for new postgraduates programmes – MSc Data for Development and the like. Just as with ICT4D, there will be a key role for practitioner hybrids – those with the ability to bridge between the world of data and the world of development – and a need for training programmes to help develop such roles. Arguably the most valuable role – to some extent trailled in my work on ICT4D 2.0 – will be the development informatics “tribrid”, that bridges the three worlds of ICT, data systems, and development.
The existing academic wateringholes and channels of development informatics will need to respond. In particular, the main ICT4D conferences and journals will need to decide whether to make a clear and strong extension of their remit into D4D. Mark Graham and I have made a first step with the 2015 IFIP WG9.4 conference in Sri Lanka; adding a “Data Revolution in International Development” track. This is an example of academic tribridisation: ensuring technology, data and development are covered in one place. It will be interesting to see what the ICTD conference series, and the main journals, do about the coming D4D wave and whether they also tribridise.
Some of the policy and practice wateringholes have already responded. One well-placed convocation is the World Telecommunication / ICT Indicators Symposium. This has, for some time, covered data, ICT and development and could grow to become a key tribrid location. More important but more difficult will be whether the WSIS follow-up process can do the same. As previously analysed, and unless it takes some decisive action, WSIS runs the risk of seeing the data-for-development bandwagon roll past it.
There are no doubt other implications of the limelight shifting from ICT4D to D4D: do add your own thoughts. These implications include value judgements. Data is not the same as technology, and the international development agenda risks taking its eye off ICT just at the moment when a digital development paradigm is emerging; a moment when ICT moves from being a tool for development to the platform for development.
Without a better connection between D4D and ICT4D we also risk losing all the lessons of the latter for the former, and turning the clock back to zero for those now entering the development informatics field riding in the data caravan. It is the privilege of those new to a field to believe they are reinventing the world. It is the burden of those experienced in a field to know they are not.Follow @CDIManchester
 “Informatics” is the complex of data, information, knowledge, information systems, and information and communication technologies.
Should we have a “95:5 rule” for ICTs and development?
Typical consumption-related uses of ICTs touch 95% of people but make only a 5% difference to their livelihoods. This covers “intensive” application of ICTs: their use to intensify an existing livelihood. Examples include use of mobiles to bring market information to farmers; access to e-government at a local kiosk, substituting a journey to district headquarters; use of a website helping handicraft producers sell their goods; or use of email by a retailer in a low-income community.
Typical production-related uses of ICTs touch 5% of people but make a 95% difference to their livelihood. This covers “extensive” application of ICTs: their use to extend the range of possible livelihoods, by created a new ICT-based livelihood. Examples include the umbrella people selling mobile phone calls by the street; or a worker from a poor community undertaking data entry work; or a mobile money service agent. So extensive ICT livelihoods only exist because of ICT and they fall into the ICT sector, broadly defined.
A classic example is the comparison of two studies from Kerala, India. The arrival of mobile phones in one fishing area led to an average 9% increase in profits for fishermen. Given 75% of income in South Indian fishing households comes from fishing, that suggests ICT consumption increased household income by 7% on average. Simultaneous to this, the Keralan government was engaged in setting up an IT impact sourcing initiative, outsourcing data entry and digitisation work to groups of women from below-poverty-line families. These new ICT jobs led to an average 75% increase in household income.
As with most quantitative findings, these specific figures don’t exactly match 5% or 95% but an overall average may get closer.
Let’s first take evidence on intensive use. Consumption-related evidence sometimes reports more than a 5% income increase. But this must be set against other work that shows a less than 5% income increase or no increase or questions the limited time-scales or scope of studies that demonstrate income increases. And it must also be set against the occasional study showing an exact match: “Internet users reported an increase of US$ 51.86 in labor income … 5.01% per year”.
Can we say that 95% of those living in the global South are digital ICT consumers? We are certainly close to that point. There were just over 90 mobile subscriptions per 100 citizens in developing countries in 2014. We need to bump that down to take account of individuals with multiple subscriptions but bump it up again to take account of shared access. The end result will be in the neighbourhood of 95%.
Turning to evidence on extensive use, many of those working in the ICT sector derive 100% of their income from their employment. We could shade that down overall given some with ICT-based livelihoods will have other income sources. The proportion of those working in the ICT sector is growing but typically less than 5% (e.g. 5.7% of employment in OECD countries but generally much lower in less-wealthy countries). As an example, India’s ICT sector represents less than 1% of India’s workforce but that must be multiplied by three given the estimate that two-thirds of India’s ICT jobs lie outside the formal ICT sector. But that estimate may exclude a number of ICT-based livelihoods, so the result may at least be heading for 5%. It is certainly increasing year-on-year.
Given these pulls in various different directions, an endpoint of 95%:5% is not unreasonable, and certainly all the evidence points to some form of strong Pareto-type distribution.
Mathematically, 5% of 95% has the same development effect as 95% of 5%. That means these two uses of ICTs should be given equal emphasis by governments, development agencies, development informatics researchers, ICT4D practitioners, etc.
But at present they are not. Intensive, consumption-related ICT application is given far, far more attention. In future that needs to be rectified, with equal emphasis given to digital inclusion by improving existing livelihoods; and to digital inclusion by creating new ICT-based livelihoods.Follow @CDIManchester
 Heeks, R. & Arun, S. (2010) Social outsourcing as a development tool: the impact of outsourcing IT services to women’s social enterprises in Kerala, Journal of International Development, 22(4), 441-454
 E.g. Aker, J.C. (2008) Does Digital Divide or Provide? The Impact of Cell Phones on Grain Markets in Niger, BREAD Working Papers (177), Bureau for Research and Economic Analysis of Development, Duke University, Durham, NC; Rizvi, S.M.H. (2011) LifeLines: livelihood solutions through mobile technology in India, in: Strengthening Rural Livelihoods, D.J. Grimshaw & S. Kala (eds), Practical Action Publishing, Rugby, UK, 53-70
 E.g. May, J., Dutton, V. & Munyakazi, L. (2011) Information and Communication Technologies as an Escape from Poverty Traps, PICTURE Africa Research Project, Nairobi; cited in Diga, K. (2013) Access and usage of ICTs by the poor, in: Connecting ICTs to Development, L. Elder, H. Emdon, R. Fuchs & B. Petrazzini (eds), Anthem Press, London, 117-136
 E.g. Srinivasan, J. & Burrell, J. (2013) Revisiting the fishers of Kerala, India, in: ICTD2013: Proceedings of the Sixth International Conference on Information and Communication Technologies and Development, J. Donner & T. Parikh (eds), 56-66
 OECD (2011) Size of the ICTsector, in: OECD Factbook 2011-2012, OECD, Paris; EC (2012) Information and Communications Technology (ICT) Sector, EU Skills Panorama, European Commission, Brussels
 NSSO (2013) Key Indicators of Employment and Unemployment in India, 2011-2012, National Sample Survey Office, Government of India, New Delhi; Nasscom (2014) India IT-BPM Overview, Nasscom, New Delhi.
 Nandi, R. (2014) Decent work and low-end IT occupation workers in Delhi, The Journal of Social Science and Humanity Research, 2(1), 9-23