Last week’s post concerned how best to measure the Information Society. This week: a look at an important new source of evidence on access and use in developing countries.
One of the most interesting sessions at last month’s Internet Governance Forum came right at the end. The attendance was disappointing – too many people had gone home, perhaps, or shopping – but the content really matters.
Why? Because it set out latest research findings on what’s really happening with access and use in developing countries on three continents: evidence that’s crucial if policies on both access and development are to tackle the complex, varied challenges involved.
This new research was funded by Canada’s International Development Research Centre (IDRC) and conducted by Research ICT Africa, LIRNEasia and Latin American research centre DIRSI. It reports on detailed, comprehensive household surveys conducted recently in 15 countries – seven in Africa, three Asia, five Latin America – with more to follow.
This is the third time these research centres have conducted household surveys like this. Earlier studies were published roughly ten and five years previously. They provide insights that aren’t available from any other source.
Why do these studies matter?
First, because the data that we have on access and use in most developing countries are really poor. The numbers used for some countries are little more than estimates based on previous years or neighbouring countries. Accuracy and timeliness are also often doubtful.
Second, because most of the data that we have’s from the supply side – from businesses and regulators (usually using business data). We need data that comes from the demand side – from users rather than suppliers. What people do with phones and Internet, rather than what they have (or what network providers think they have). Few developing countries have reliable household surveys that can provide this, and those that have conduct them rarely and in little detail.
Third, the access model that’s most prevalent in developing countries – prepaid mobile – makes demand side data even more important. Supply-side data on subscriptions, not subscribers, exaggerate the number of people who’re online. They’re very little help if we want to disaggregate by gender, or in different age groups. Disaggregation’s crucial for planning both business needs and applications that are focused on development.
Fourth, because those demand side studies that have been published from time to time have usually been too narrow, looking at a few places at one point in time. Samples have been too small, or poorly representative. Some come from organisations with a business model to promote. More attention’s been given to publicising them than to the methodology and analytical rigour that are essential if policymakers are to be informed, not merely influenced.
And context matters
Fifth, because the experience of different groups within society is different. Almost every country experiences annual improvements in uptake of mobile telephony, broadband and Internet. But those overall improvements aren’t evenly distributed. Gains may be made only in urban areas, or by the rich; by men rather than by women, and so forth.
Sixth, because countries also differ greatly. Too much international policy gets made for the developing world in general rather than the specific contexts of individual countries. But circumstances can be very different, even in neighbouring countries, as these latest surveys show.
What does the research say?
So these studies are important. They’re based on larger samples than most other household surveys, and their methodology is rigorous. They’re very detailed and the same approach has been taken in a significant number of countries, which can be compared with one another. It will take a while to analyse them fully, but their data will be open to researchers.
So far, only headlines are available, and I’ve just space to point at six findings that should make ICT advocates and policymakers sit and think. Six points to ponder next time you hear someone talk about ICTs and development as if everywhere’s the same and everyone’s experience alike. I’ll come back to these (and other) findings from these surveys later in the year.
First, there are big differences in take-up between and within regions. All five Latin American countries in the survey have higher rates of Internet use than any of the seven surveyed in Africa. User rates on that continent vary greatly, from fifty per cent (South Africa) to less than ten (Rwanda).
Second, there are big gaps between urban and rural areas in most countries. Take Tanzania for example. Urban users have, on average, been online four years; those in rural areas for two. Ten per cent of phone users have smartphones in rural areas; almost forty in the towns and cities. Thirty four per cent of urban households in India, meanwhile, have Internet access, but just fourteen in rural areas. This reflects and potentially exacerbates urban/rural divides.
Third, the gender digital divide is real andsubstantial. It’s particularly acute in South Asia. This survey finds a 46% gap between rates of male and female ownership of mobile phones in India; an even higher gap (57%) in rates of Internet use. The gender gaps are smaller in Africa and Latin America, and occasionally reversed. It's clear that these gaps reflect structural inequalities of education, income and opportunity between women and men.
Fourth, countries with high profiles in ICT4D aren’t necessarily performing better. Few countries in Africa have invested as heavily in online infrastructure and services as Rwanda, or had so many development programmes targeted on ICTs. Yet these figures show that it has lower mobile access rates than all but one of the other African countries surveyed; and lower smartphone and social media usage rates than all those peers.
Fifth, the ‘killer app’, as it’s been called, is social networking. Very high percentages of those online, in these surveys, are using Facebook. Indeed, some who use Facebook don’t think they’re using Internet. Indian respondents put social media, chatting and entertainment at the top of the list of time spent on the Internet. This is not surprising, but its implications – from the value of zero-rating to designing developmental applications – are too often overlooked.
And sixth, while reasons given for not using the Internet vary substantially, a lot of people still have low awareness of it. Two thirds of those surveyed in Bangladesh did not know about the Internet; two thirds of those who didn’t own a smartphone said it was because they didn’t need one. The Internet may look more pervasive and more widespread than it is because its non-users are less visible to bureaucrats, to business leaders, and to advocates of ICT4D.
These are but tiny samples from the data in these surveys. The quality and richness of what’s in them means they should be widely read and should have considerable potential impact. If policymakers are to make effective use of ICTs they need to know what’s actually happening in their societies, rather than relying on a mix of old or unreliable data, hype and glib assumptions. They need surveys like these and they need to take account of what they say.
So more studies are needed, in more countries, and they need to be more regular if they’re to capture the rapid changes that are taking place. It would be good if comprehensive studies like these could be done in sample countries every three years rather than five. Good, too, if more countries could be covered. But that’s going to require more funders. IDRC has done a splendid job but it can only finance studies in a few countries every half-decade or so.
And it would be good if questions concerning ICTs could be included in regular surveys that are held by governments in all countries for other purposes – not just censuses, which are infrequent, but more regular surveys such as those of household income and expenditure wherever these take place.
Demand side data of this kind add greatly to the picture that we have of changing ICT access and use, but they can’t tell us everything, any more than the supply-side data can. They also have their limits and distortions. (Respondents are inclined to stress uses that they think are to their credit (such as education), for example, not those that they think don’t.
We need accurate and timely data from both supply and demand sides of the market – from network operators and online service providers as well as household surveys. Together, these will tell us more than either will alone. More openness from networks and online businesses would help to bring about that goal.
Next week: some thoughts about children online.