Hand over your data!

Researchers are robbing the poor by keeping data out of the public domain, argues David Hulme

Researchers from European and US universities are in effect stealing from the poor by holding on to data about poverty.

The Chronic Poverty Research Centre - a partnership of universities, research institutes and advocacy non-governmental organisations from the UK, Africa and South Asia - has found evidence that researchers and their agencies "capture" datasets about poverty and wellbeing, hanging on to them for years on end. This may enhance academic reputations, but it also slows the understanding of poverty and hampers poverty-reduction efforts.

Panel datasets produced by repeated questionnaire surveys of the same households can be used to analyse changes in the livelihoods of poor households in great detail. But they are rare in developing countries because they are demanding in terms of money and management. New datasets are needed, but much better use could be made of existing information if it were made available to more researchers more quickly.

Three particular types of "data stealing" have concerned CPRC researchers.

In universities and research institutes in Europe and the US, leading analysts hold on to some quantitative datasets, gaining monopoly use of them for extended periods. Local poverty researchers in African and Asian countries report that they cannot access crucial statistics about what is happening in their own nations while researchers at agencies in Washington DC have unlimited access to the necessary figures.

A different problem arises in South Asian countries when official statistical agencies refuse to make data available for public use - even to paying clients - for years after collection. Sometimes this is because of red tape, but other times it is because statisticians in these agencies can accept private contracts to analyse datasets. The longer the information stays outside the public domain, the more contracts they can accept.

Another issue is the reluctance of qualitative researchers to make their data, such as life histories and open interviews, publicly available, although the internet now makes this easy.

As most of these datasets are collected with the aid of public funds, they should not be converted into private property by elite researchers.

Moreover, restricting public access to datasets weakens the scientific base of research on poverty and development. It prevents other researchers from testing the results of those who have done the initial analysis, and it reduces the likelihood of identifying effective policies by slowing the advancement of useful knowledge. When other researchers eventually get access to the data, which may be many years later, their analyses may have little relevance to policy because of the time that has lapsed.

Finally, although interviewers, researchers and research fund staff are on a salary, the main "input" in such datasets is the time of poor people in developing countries answering questions, for which they are unpaid.

What should be done? First, research funding bodies should set dates for the publication of datasets. It is not unreasonable for researchers who design surveys to have a period of exclusive analysis after data have been cleaned. But this period should not be more than 12 months. Some agencies, such as the UK's Economic and Social Research Council, require that data be made available, but foreign aid agencies and foundations often have no such procedures.

In addition, funding bodies may need to give grant recipients extra money to pay the costs of advising other analysts about their data. Researchers should also monitor each other, and development NGOs and advocacy groups should name and shame researchers and agencies who delay data entering the public domain.

Good practice is emerging. The International Food Policy Research Institute puts its datasets on its website. Researchers at the Programme for Land and Agrarian Studies at South Africa's University of the Western Cape are committed to placing their panel datasets on poverty on the web six months after they have been cleaned. The Bangladesh Institute of Development Studies has created a website that aims to consolidate all the poverty datasets on the country at a single location free of charge. It is time for other researchers and their institutions to follow these examples.

David Hulme is director of the Chronic Poverty Research Centre and professor, Institute for Development Policy and Management, University of Manchester. A version of this article appeared in id21 insights, the bi-monthly review of international development research ( www.id21.org ) The CPRC will host an international conference on poverty and development policy on April 7-9. Details at www.chronicpoverty.org