Abstract
Overview. In 2008-2009, the Bolivian government announced its intention to build a road through the lands of indigenous peoples in the tropical lowlands of the department of Beni. Using a natural experimental approach, the goal of this study was to measure changes in well-being of indigenous people by comparing villages affected and unaffected by the road, before and after road construction. Bolivia did not build the road due to protests. To tests hypotheses, a broad range of measures were taken from the same adults and households in 2012 and 2013. Because the road was not built, we could not estimate impacts, but we used the data to study attitudes toward road construction and assess which children are left behind in Bolivia’s conditional cash transfer programs for schooling. Researchers and policy-makers have debated the human and ecological costs of road construction through the Amazon and portrayed native Amazonians as Noble Savages keen on keeping outsiders at arms’ length. We found most indigenous people favored the road (Reyes-García et al. 2020a). Data were used to analyze a conditional cash transfer program for schooling among four native Amazonian groups (Bauchet et al. 2018). We found that Tsimane’ children were less likely to enroll in such programs.
Publications. Three publications came out of the research:
Bauchet, J.; Undurraga, E.A.; Reyes-García, V.; Behrman, R.; Godoy, R. 2018. Conditional cash transfers for primary education: Which children are left out? World Development 105:1-12.
Reyes-García, V.; Fernández-Llamazares, A.; Bauchet, J.; Godoy, R. 2020. Variety of indigenous peoples’ opinions of large infrastructure projects: The TIPNIS road in the Bolivian Amazon. World Development 127: e104751.
Reyes-García, V., Andrés-Conejero, O., Fernández-Llamazares, Á., Díaz-Reviriego, I., Molina, JL. 2019. A road to conflict: Stakeholder’s and social network analysis of the media portrayals of a social-environmental conflict in Bolivia. Society and Natural Resources. 32(4):452-472.
The datasets. We are making available to the public two types of de-identified datasets in Stata 17:
[a] A complete, appended and merged dataset of individuals, households, and villages. The datasets have been cleaned to the best of our ability. Variables are defined in the dataset itself.
[b] Four ancillary, stand-alone datasets: 1) code for animals, 2) code for plants, 3) reasons for leaving or moving into the communities, and 4) prices of selected items in two towns in the study area. Datasets 1-2 are useful to identify plants and animals in some of the modules. For example, in a question we might have asked “What animal did you hunt in the past week?” and the answer was coded as either “Deer” or with a number, like 234. In dataset 1) the user will be able to find that code 234 stands for deer. Dataset 3) was collected at the village level and allowed for multiple observations for a village in a year; for example, in a community survey, community leaders might have said that people left the village for reasons A, B, and C, each of which would be entered as a discrete row. So constructed, the data could not be easily merged with the other datasets. Dataset 4) might be useful for those wishing to value asset wealth or consumption when the village surveys on prices had meek information. In both years, surveyors asked about prices for a basket of goods at the village gate, but often there were missing values because villagers did not buy or sell the good. When this happens, town prices might be used to impute values to missing observations.
A dataset on inter-community relation is not shown because it contained the names of villagers or officials.
Other. We also include the surveys in Spanish and the proposal to the Program of Cultural Anthropology of the National Science Foundation (The effects of roads on indigenous people’s well-being and use of natural resources. A natural experiment in lowland Bolivia; PI: Godoy; NSF #: 0963999; $350,000. 2010-14). Survey data was entered into Microsoft Access, from which individual files were exported to Stata 17 before appending, merging, and cleaning. The coarse, de-identified Access datasets will be placed at the electronic library of Brandeis University.
Date: March 8, 2022
The clean dataset of this study was submitted to ICPSR for review and storage but at the time of this writing (November 2022) has not yet been approved and made available to the public.