Living Standard Measurement Study: When Data Collection is Intrepid

• Living Standard Measurement Study (LSMS) team plays vital role and travels many back roads.
• Eighty-eight surveys of very poor households spanning 26 years, 90% of them open to the public.
• Firsthand microdata helps people understand living standards, poverty and inequality

For two to three weeks at a time, Kathleen Beegle disappears into the wide open spaces of rural Africa. She often lands in a plane on a dirt runway, rents a pick-up truck and drives with a colleague for four to six hours to her destinations.

That is, if they are lucky. Sometimes, they are foiled by a washed-out bridge or impassibly muddy roads. Then, it’s more hours of bone-jarring driving. A handful of religious music stations on the car radio, along with old 1980s cassettes of Dolly Parton and Julio Iglesias, help pass the time. At night, Beegle often sleeps in the car or under a mosquito net. Most villages don’t yet have electricity, she says. “It’s so dark you see every star.”

Beegle isn’t an explorer for the National Geographic. She is an economist and one of nine members of the household survey program, the Living Standard Measurement Study, in the World Bank’s Development Research Group. Team members travel far to collect data about the world’s poorest households, analyze them and conduct research on survey methods.

LSMS data play a critical role in helping policy makers, the development community and others understand living standards, poverty and inequality in a particular country, region or the world. Over the past 26 years, the team has helped conduct 88 surveys, from Nicaragua to Timor Leste to Iraq.

The LSMS surveys are done at the request of governments, donors or World Bank country teams. As part of an initiative funded by the Bill and Melinda Gates Foundation, new surveys are under way in six countries in sub-Saharan Africa to improve agricultural data in the region, which are currently unreliable or completely lacking. As with other LSMS surveys, the money is funneled through the World Bank to those countries – mostly to national statistical offices – to collect data with the LSMS team’s technical assistance.

For 30 years, the LSMS team has been working with individual governments to ensure open access to LSMS data, which is owned by individual governments, says LSMS manager Kinnon Scott. For example, as a condition to participate in LSMS programs in places such as sub-Saharan Africa and Latin America, governments agreed in advance to make their LSMS data public. So far, results for 90% of all LSMS surveys are accessible to the public. Most can be downloaded from the World Bank’s LSMS website, www.worldbank.org/lsms, but some are also distributed on CDs by mail or directly from national statistical offices, says Diane Steele, LSMS household survey coordinator.

From start to finish, the LSMS team works with government institutions – especially their national statistical offices – to design, implement and analyze the surveys. The team focuses on teaching the know-how, the skills needed to build and sustain a data-collection system unique to a particular country, to staff and field workers. They also engage policy makers, research communities and local counterparts.

The approach has paid off. For example, Peru, where the first LSMS survey was conducted in 1985, has included the survey in the government’s regular data-collection effort – without any additional technical advice from the LSMS team. Nepal, Albania and other countries also have conducted surveys on their own. Many more surveys around the world have adopted LSMS methodologies and practices.

LSMS survey questionnaires cover a range of topics, from demographics, to education, health, labor, consumption, finance, farm production, and non-farm activities. Each questionnaire addresses context-specific policy questions, reflecting data demand from across the government. For example, migration is important to Albania, fertilizer use matters in Malawi and conditional cash transfers are of interest in Panama. The questionnaires are compatible enough to allow many comparisons between surveys.

Some questions are trickier than others: how do you convert “four buckets of corn,” “a heap of tomatoes” or “three plates of cereal” into measurable units of quantities or calories? A plateful of cashews with a flat surface may be 300 grams, but a full plate could double that calorie count. In addition, some questions are difficult to translate into local languages, and the team conducts extensive pilot testing and fieldwork in advance to make the questionnaires work in each specific context.

“A survey is just as much an art as it is a science,” says Gero Carletto, manager of the project on agricultural data in Africa. “Correcting small mistakes in survey design and implementation, and catching them early on, can have huge payoffs in terms of data quality.”

The LSMS staff spends several months to a year on the road, meeting with officials and development partners, but mostly working closely with national statistical offices and local teams to help ensure the best possible data quality. Indeed, the team is known for the development and adoption of strict data-quality control protocols.

Beegle and the other members of the team value their time traveling with local interview teams. The entourage, which usually includes a supervisor and three or four interviewers, often pack measuring equipment, luggage, cooking pots, and water jugs onto the roof of their sports-utility vehicle. Their trips can last three or four months – or sometimes a year. An LSMS team member might join them for a few days or a few weeks.

Finding a representative, random sample of households and individuals can be a challenge in rural Africa or South Asia, where streets often aren’t marked. And, to talk to working members of the family, interviewers often have to visit the house very early in the morning or late at night.
Some surveys are especially challenging. The Gates-funded project in rural Africa, for example, involves panel data for as many as 5,000 households per country, which will allow comparisons of the same people over time. But in between surveys, people die, marry, move hundreds of miles away, wind up in jail, or even go into hiding from creditors and others.

But once they find the household to interview, people are often quite happy to open up. And overtime, Beegle says, you know so much about a family that you almost feel like part of their lives. “They are often so nice. They offer you tea, give you bags of ground nuts. They are really poor people and they are so generous and offer me what they have,” she says.