Amazon study shows accuracy of native-collected data

Oct. 17, 2011, 2:03 a.m.

Stanford researchers led by ecologist and visiting scholar in biology Jose Fragoso have conducted a five-year study in the Amazon Basin demonstrating that data collected by native peoples trained by researchers is equally as accurate as data collected by researchers themselves.

The researchers, who set out to determine the effects of humans on vertebrate populations in a 48,000-square-kilometer region, soon found that researchers themselves could not collect all the data.

“The only way you are going to understand what is in the Amazon in terms of plants and animals and the environment is to use this approach of training indigenous and the other local people to work with scientists,” Fragoso told the Stanford Report. “If I had tried to use only scientists, postdocs and graduate students to do the work, it would not have been accomplished.”

Fragoso’s team worked with the Makushi and Wapishana peoples of the Rupununi region in Guyana. They trained 340 individuals in 28 villages to collect consistent, systematic data. Older villagers, many of whom were illiterate or unable to do arithmetic, were paired with younger ones who had these skills but did not know the area’s animals and plants as well.

“This is the first study at a really large scale that shows that consistently valid field data can be collected by trained indigenous peoples, and it can be done really well,” Fragoso said. “We have measured the error and discovered that 28 percent of villages experienced some data fabrication. This originated from about 5 percent with 18 out of 335 of technicians fabricating data, which may not be much different than what occurs in the community of scientists.”

This statistical analysis to confirm data accuracy was done by additional collection sessions by a different team of technicians or the researchers themselves. The team found that the most accurate data was collected in communities with strong leadership or those that were part of an association of villages; fabricated data came from those with less oversight.

“The indigenous technicians are no more corrupt, sloppy or lazy than we are,” Fragoso said.

Their results are published in the October issue of BioScience. The project was funded by the National Science Foundation and the Gordon and Betty Moore Foundation.

“I have presented this study to some pretty unreceptive groups, such as at scientific meetings, but by the end of the presentation, audience members are either convinced or at least they doubt their argument, which is a major achievement in itself,” he said. “One thing about the scientific community–if you have enough solid data and the analysis is well done, there is very little you can argue against.”

– Ellora Israni

Login or create an account

JOIN THE STANFORD DAILY

application deadline
Friday, Oct. 11

Days
Hours
Minutes
Seconds