Measure2003

10.1177/1525822X03254847 ARTICLEFIELD METHODSVadez et al. / SELF-REPORTS IN MEASURING DEFORESTATION
Validity of Self-Reports to
Measure Deforestation: Evidence
from the Bolivian Lowlands
VINCENT VADEZ
Brandeis University
VICTORIA REYES-GARCÍA
University of Florida
RICARDO GODOY
Brandeis University
LUKE WILLIAMS
LILIAN APAZA
Universidad Mayor de San Andrés
ELIZABETH BYRON
TOMÁS HUANCA
Beni, Bolivia
WILLIAM R. LEONARD
Northwestern University
EDDY PÉREZ
Universidad Mayor de San Andrés
DAVID WILKIE
Wildlife Conservation Society
To assess rates of deforestation, researchers typically use questionnaires. But do
questionnaires provide accurate information about the extent of forest clearance by
households? In this article, the authors provide data on the amount of deforestation
in a Tsimane’ Amerindian village (Bolivia) and assess informant error by cross-
checking three different assessments: (1) a direct physical measure by a research
team of each plot cleared from the forest, (2) an estimate by the household head of the
entire area cleared by his household during the year before the interview, and (3) an
estimate by the plot owners of the area cleared of each plot he owns. Results show a
high correlation between direct measures and estimates of areas provided by infor-
mants; plot owners provided more accurate information than heads of households.
Field Methods, Vol. 15, No. 3, August 2003 289–304
DOI: 10.1177/1525822X03254847
© 2003 Sage Publications
289

Results suggest that asking Amerindians about the area of forest cleared will yield
reliable estimates of deforestation.
Keywords: deforestation; indigenous people; informant accuracy; slash-and-burn
agriculture; education; Tsimane’ Amerindians; Bolivia
Throughout the world, the area under forest cover has fallen sharply and
continues to decline (Mathews 2000; Nepstad et al. 2001). The main causes
of deforestation include slash-and-burn agriculture, logging, and cattle
ranching (Hecht 1998; Demmer and Overman 1999). Indigenous people
practice small-scale, slash-and-burn agriculture. Although some scholars
think indigenous people do not deforest much (Pimentel and Heichel 1991;
Myers 1992), others (Godoy 2001) say that indigenous people will contribute
to higher levels of forest clearance in the future. Population pressure (Picchi
1991) and increasing integration into a market economy might induce house-
holds to expand the area of forest cleared (Simmons 1997; Godoy 2001). If
the trend holds, the amount of forest cleared by indigenous people has to be
measured, and the determinants of deforestation need to be analyzed before
one can propose solutions. So far, there have been few attempts to measure
the amount of forest cleared by indigenous people (Godoy et al. 1997).
Before estimating the area deforested, one must define deforestation and
the type of forest it refers to. Deforestation is typically defined as the tempo-
ral or permanent clearance of forest vegetation (Lanly 1982; Grainger 1993),
but researchers generally make no distinction between old-growth forest and
fallow forest. Fallow forest refers to secondary forest (Myers 1992; Finegan
1996; Silver, Brown, and Lugo 1996; Smith et al. 1999). The earliest assess-
ments of deforestation by the U.N. Food and Agricultural Organization
(FAO) do not distinguish between fallow and old-growth forests (FAO
1993). Since indigenous people usually cultivate in a radius of 1–2 km from
their residence, they tend to clear fallow forest instead of relocating close to
old-growth forest. To measure total deforestation by indigenous people, we
have to measure both old-growth and fallow forest cleared.
Since indigenous people live in limited, often circumscribed territory and
move a great deal in search of fish, game, and farmlands (Reyes-García
2001), satellite imagery would provide only a rough estimate of deforested
area. To get household measures of deforestation, we can draw on several
approaches. First, we can get survey data on the amount of forest cleared by a
household and on pertinent covariates likely to affect the amount of forest
cleared. Before we can draw reliable inferences from survey data, we need to
know the direction and magnitude of biases from measurement errors of
informants (Bernard et al. 1984; Romney, Weller, and Batchelder 1986).
290 FIELD METHODS

Second, we can make direct measurements of deforestation using a measur-
ing tape and a compass. Direct measures provide more accurate estimates of
deforestation but require more time. Last, we can calibrate the information
from surveys with direct measures made in the field.
If we intend to calibrate the information on deforestation from surveys, we
should also take into account factors likely to affect measurement errors in
surveys. For example, formal education might affect the size and the direc-
tion of errors when estimating field size. People knowing the basics of arith-
metic might make smaller errors when estimating the size of their fields
because they are more adept at computations. Indigenous people are often
illiterate, do not measure their plot, and are more likely to estimate the area
they clear with error.
Here, we analyze data on the area of forest cleared by households in one
village of Tsimane’ Amerindians in the tropical rain forest of Bolivia. We
distinguish between areas cleared from fallow forest and from old-growth
forest and estimate how much error stems from informant answers in estimat-
ing (1) the area of forest cleared and (2) the type of forest (old growth or fal-
low) cleared. We do this by comparing the following three estimates: (1) the
area cleared by the household as estimated by the household head, (2) the
area cleared in each plot as estimated by the owner of the plot, and (3) the area
of forest cleared by the household or the area of the plot as measured with a
tape and a compass by a trained research team. This study builds on a previ-
ous study of error in estimation of field size in Honduras (Godoy et al. 1998)
by focusing on different methods for judging accuracy. The previous study
focused more on the socioeconomic covariates of error.
THE TSIMANE’
The Tsimane’ are a foraging and horticulturalist society of approximately
seven thousand people living mostly along the Maniqui and Apere rivers in
the department of Beni in the Bolivian lowlands. Ellis (1996), Chicchón
(1992), Huanca (1999), and Reyes-García (2001) provided ethnographic
descriptions of the Tsimane’.
The Tsimane’ practice traditional slash-and-burn horticulture. Between
June and September, they clear forests using simple tools. The type of crop
they intend to plant affects the decision of how much forest to clear. When
deciding whether and how much fallow forest to clear, the Tsimane’ also take
into account the size of the tree trunks rather than the number of years that
have elapsed since the last cropping cycle. The Tsimane’ do not generally
measure plots. Once cleared, they plant their main staples—rice, maize, cas-
Vadez et al. / SELF-REPORTS IN MEASURING DEFORESTATION 291

sava, and plantain—and crops of less importance (e.g., sugarcane, ground-
nuts, or sweet potatoes). After a short cultivation cycle of only one to two
years, they abandon the plot to clear another plot. The abandoned field
remains an important purveyor of game, fruits, and plants long after it has
ceased to produce farm crops and has managed to produce many plants (e.g.,
shafts for arrows, citrus) (Huanca 1999). Once the plot is abandoned, forest
vegetation takes over. Years later, it is sometimes hard to notice whether or
not a forest patch had been previously cleared. During the past thirty years,
under the influence of Protestant missionaries, the Tsimane’ have progres-
sively modified their settlement and migratory patterns, from extended fami-
lies moving along the river to find farmland and game, to more sedentary set-
tlements in small villages of ten to thirty households. Households in more
sedentary villages find it increasingly necessary to clear fallow forest.
METHOD
Setting
Fieldwork took place in the Tsimane’ village of Yaranda (15′16.369 S,
66′50.838 W) along the Maniqui River. We selected the village because it is
representative of other Tsimane’ communities in access, size, type of subsis-
tence, and degree of integration to the market economy. People still have old-
growth forest to clear within walking distance from their homes. The com-
munity has twenty-five households, a school, and two teachers, who cover
grades one through five. The village lacks access by road. During the dry
summer months, between June and November, people reach the nearest mar-
ket town by walking for two days or by canoeing for three days. Walking is
possible during the wet month, but people are reluctant to visit the market
towns canoeing because of the difficulties of paddling upriver.
Sample
The sample consists of twenty-five households and all the plots (n = 36)
they cleared during 1999.
Education
Since educational level and skills in arithmetic might affect the error in
estimation of field size, we also asked participants about the maximum edu-
cation level they had reached. To measure skills in arithmetic, we gave a test
in which we asked individuals four questions that required them to add, sub-
292 FIELD METHODS

tract, multiply, and divide. We assigned one point to each correct answer; the
total score could range from zero to four.
Plot Measures
Our objective was to estimate and compare the size of plots cleared from
old-growth forest or fallow forests using the following three methods.
Method A: direct measure. A non-Tsimane’ researcher and an adult
Tsimane’ guide measured all the plots (n = 36) in the community during one
growing season (September and November 1999). To decide whether plots
were from old-growth or fallow forests, we had a Tsimane’ informant tell us
the type of forest from which each plot had been cleared. When we arrived at
a plot, we asked the Tsimane’ informant to tell us the name of the person who
had cleared the plot. We measured the plot sides with a measuring tape and
the angles of the plots with a compass. We entered the information into a
computer program that drew the field and calculated the surface area and the
perimeter. We remeasured plots if the calculation error given by the com-
puter program was more than 5% of the plot area. Since the measurement
consisted of measuring the sides of the fields and the angles between sides,
this could result in the perimeter not reaching closure. In that case, the com-
puter program provided an estimate of the gap or open area of the perimeter.
For the analysis, we used the direct measure of the plot by the research team
as the most accurate estimate of the true area.
Method B: interview of the household head. In May 2000, about four
months after having measured the field, we interviewed the male household
heads (n = 24) without referring to the information already gathered through
method A. We asked the household heads to estimate the total area of forest
they had cleared that year; we did not ask them about the area cleared for spe-
cific fields. We recorded the estimates in tareas (1 tarea = 0.1 ha), the unit of
measure Tsimane’ use to estimate area.
Method C: interview of the plot owner. A month later, in June, we con-
ducted another interview with the person who had cleared the plot, and we
did this for every plot measured with method A. To remind the plot cutter
what plot we wanted him to estimate, we referred to the physical location of
the plot and to the crop that he had planted on that plot at the beginning of the
season. We interviewed any person (not just the household head) who had
cleared a plot. For each of the plots measured with method A, we asked the
person who had cleared the plot to tell us the type of forest cleared and to esti-

mate the surface of the plot cleared. In many cases, people did not estimate
the surface in tareas but gave an estimate of the length of the field sides in
meters. In that case, we assumed the fields were perfect rectangles and con-
verted the meter data into tareas. We used methods B and C one month apart
to avoid leaks of information among informants and to reduce possible con-
fusion between the two questions.
RESULTS AND DISCUSSION
Summary Statistics
When we measured plots (method A), we asked the Tsimane’ guide to
decide whether the plot had been cleared from old-growth or fallow forest.
When we interviewed the plot cutter (method C), we also asked the cutter
whether the plot had been cleared from old-growth or fallow forest. The
results show high correlation between what the plot cutter reported and what
the Tsimane’ guide had said; in only five out of the thirty-six plots did the plot
cutter and the Tsimane’ guide differ in their identification of forest type.
During 2000, the entire village cleared thirty-six plots covering an area of
15.36 ha. The average household had 1.5 plots. The mean area cleared per
household was 0.694 ha, and mean plot area was 0.427 ha (see Table 1).
People cleared about the same number of fields from old-growth (nine-
teen) and from fallow forests (seventeen). The surface cleared from old-
growth forest represented about two-thirds of the total area cleared (see
Table 1). Plots cleared from old-growth forest were larger (0.52 ha/house-
hold) than those cleared from fallow forest (0.32 ha/household).
Relation between Measured Area Cleared by
Household and Estimate by Household Head
To estimate the total area of forest cleared by a household, we added the
area cleared in each plot by the household as determined by the research team
(method A). We regressed the values measured by the team against the area
estimated by the household head. With perfect accuracy, we would expect a
slope equal to 1. The regression yielded a slope of 0.51 (t = 3.56, p < .002, R2
=
.40; see Figure 1), far from the expected value of 1. The results suggest that
male heads of households (method B) provide inaccurate estimates of the
total amount of area deforested by the household. Household heads underes-
timate the area of forest cleared. Later we explore possible causes of
underestimation.
294 FIELD METHODS

Relation between Measured Area of Plot and Estimate by Plot Owner
Because the household head provides a biased estimation of the forest
area cleared by the household, we next compare the area measured of each
plot with the estimation provided by the owner of the plot. We regressed the
values measured by the researcher team against the estimate given by the plot
owner. We found a slope of 0.84 (t = 8.60, p < .0001, R2
= .68; see Figure 2).
We tested that the coefficient was equal to 1 and accepted the null hypothesis
that the coefficient was equal to 1 at the 95% confidence interval (F = 3.55;
p < .0713), suggesting that plot cutters provide accurate estimates of the size
of their plots.
Relation between Measured Area Cleared by
Household and Estimation Error by Household Head
Household heads may have underestimated the total forest area cleared by
their household if they forgot about plots cleared by other members of their
household. If true, we would expect to find a positive correlation between the
error of estimation made by the household head and the actual area cleared
per household as measured by the researcher team. Figure 3 shows the slope
of the regression of estimation error made by the household head (dependent
variable) and the area of forest actually cleared by the household. As the fig-
ure illustrates, we found a statistically significant positive correlation (slope =
0.49, t = 3.39, p < .003, R2
= .38). In Figure 3, points along the line y = 0 imply
small informant error. The pattern of Figure 3 suggests that household heads
TABLE 1
Summary Statistics of Cleared Plots by Method of Estimation
Forest Type Total Area
(number per Forest Total Area Area
of plots) Type (ha) Area (ha) HH
–1
(ha) Plot
–1
(ha)
Method OGF FF OGF FF
A 19 17 9.83 5.53 15.36 0.694 ± 0.059 0.427 ± 0.04
B NA NA NA NA 13.13 0.597 ± 0.069 —
C 20 16 9.71 4.36 14.07 — 0.391 ± 0.041
NOTE: Method A = plot area measured by researcher’s team); method B = household area
cleared estimated by household head (HH); method C = plot area cleared estimated by plot
owner; NA = not asked. Plot type refers to the forest type cleared for agriculture, either old-
growth forest (OGF) or fallow forest (FF). Plot surface is given in tareas (1 tarea = 0.1 ha). Mean
area cleared given per household (area HH–1
) or per plot (area plot
–1
).

are more likely to provide faulty estimations when the true area cleared
exceeds 4 tareas. In households with larger areas of cleared land, household
heads are more likely to forget plots.
Relation between Measured Area of
Plot and Estimation Error by Plot Owner
We regressed the error in estimation by the owner of the plot against the
area measured by the research team (see Figure 4) and found no significant
relation (coefficient = .08, t = 0.93, p < .36, R2
= .02); the slope was not signif-
icantly different from 0. The results suggest that the error in estimation by the
owners of the plot is not related to the size of the plot.
We also expressed the error of estimation by plot owner as a percentage of
the measured area of the plot (area estimated by plot owner minus area mea-
296 FIELD METHODS
Tareas per household (measured)
2 4 6 8 10 12
Tareasperhousehold(estimated)
0
2
4
6
8
y=2.0 + 0.51x R2=0.40 (n=21)
FIGURE 1
Plot between the Estimates of Area Cleared by Households,
Given by Each Household Head, and the Actual Values of
Area Cleared of Households, Measured by an Independent Researcher
NOTE: Data are expressed in tareas (1 tarea = 0.1 ha; n = 21).

sured by the team divided by the area measured by the team). Because plot
owners either overestimate or underestimate the size of their fields, the dif-
ference between the area measured by the researcher team and the estimate
given by the plot owner can either be negative or positive; we therefore took
the absolute values. We regressed the absolute values against the area mea-
sured by the research team (see Figure 5). We found a hyperbolic relation:
For plots smaller than about 3 tareas, plot owners made an error that repre-
sented a high percentage of the plot area measured. This might be due to
informants’ using relatively coarse units of area in their responses no matter
how large or small their plots were. Figure 2 shows that the error seemed to be
rather constant across the range of field sizes. The heaping of the curve would
Plot area in tareas (measured)
0 1 2 3 4 5 6 7 8 9
Plotareaintareas(estimated)
0
2
4
6
8
10
y=0.3 + 0.84x R
2
=0.68 (n=36)
FIGURE 2
Plot between the Estimates of Area Cleared of Plots,
Given by Each Plot Owner, and Actual Values of Area Cleared
of Plots, Measured by an Independent Researcher
NOTE: Data are expressed in tareas (1 tarea = 0.1 ha; n = 36).

be explained by errors representing a large proportion of the field size in
cases of small fields; the proportion decreases as field size increases.
Since plots from fallow forests were usually smaller (3.2 tareas) than plots
from old-growth forests (5.2 tareas) (calculated from Table 1), the results are
consistent with the fact that there was more error in the estimation of fallow
forest (total of 5.53 ha cleared measured against 4.36 ha estimated; see
Table 1) than in the estimation of old-growth forest (total of 9.83 ha cleared
measured against 9.71 ha estimated; see Table 1).
Relation between Human Capital and
Estimation Error by Plot Owner
We also estimated the effect of education on the error made by the plot
owner. The Pearson correlation coefficients between error, school grades,
298 FIELD METHODS
Tareas per household (measured)
2 4 6 8 10 12
Estimationerrorbyhouseholdhead
(tareas)
-3
-2
-1
0
1
2
3
4
5
6
y=-2.0 + 0.49x R2=0.38 (n=21)
FIGURE 3
Plot between the Estimation Error by the Household Head
Minus the Estimate Given by the Household Head and the
Area Cleared of Households, Measured by an Independent Researcher
NOTE: Estimation error by the household head is the actual value of area cleared by the house-
hold (measured by an independent researcher). Data are expressed in tareas (1 tarea = 0.1 ha; n =
21).

and arithmetic skills were –.35 (p < .04) and –.01 (p < .94). The Pearson cor-
relation coefficient between school grades and arithmetic skill was .53 (p <
.0009).Averagenumberofschoolgradesachievedbyinformantwas2.2(SD=
2.6), average mark on the arithmetic test was 2.0 (SD = 1.6), and average
error made (in tareas) was 1.05 (SD = 1.02). We took the absolute values of
the estimation error by the plot owner and ran a multiple regression with
skills in arithmetic and the number of school grades completed by the infor-
mant as explanatory variables. We found that education bore a positive corre-
lation with the error made by the plot owner. Having one more year of school-
ing correlated with lower estimation errors of 0.21 tareas (t = –2.77, p < .001;
see Table 2). In contrast, an increase of one point in the arithmetic test had no
significant effect on the estimation error (t = 1.41, p < .17).
0 1 2 3 4 5 6 7 8 9
Estimationerrorbyplotowner
(tareas)
-4
-3
-2
-1
0
1
2
3
4
5
y = -0.11 + 0.08x R
2
=0.02 (n=36)
FIGURE 4
Plot between the Estimation Error by Plot Owner
Minus the Estimate Given by the Plot Owner and the Area
Cleared of Plot, Measured by an Independent Researcher
NOTE: The estimation error by plot owner is the actual value measured of area cleared of plot
(measuredbyanindependentresearcher).Dataareexpressedintareas(1tarea=0.1ha; n=36).

CONCLUSION
The main empirical finding of this study is that informants generally give
accurate estimates of the area of forest they clear, provided they are asked
about the area they cleared on their own plots. In contrast, household heads
asked to estimate the area cleared by their entire household underestimate the
area cleared by their entire household. Household heads are likely to forget to
include the area cleared by other members of their households. When they
commit errors in estimations, informants tend to underestimate the true area
cleared.
The educational level of the plot owner had an important weight on his
estimation error, with more educated men making more accurate estima-
tions. Previous studies show that among lowland Amerindians, schooling
300 FIELD METHODS
0 1 2 3 4 5 6 7 8 9
Absolutevalueofestimationerrorbyplotowner
(inpercentageofmeasuredarea)
0
20
40
60
80
100
120
140
160
FIGURE 5
Plot between the Absolute Values of Estimation Errors by
Plot Owner, Expressed as a Percentage of the Actual Plot Area
(Measured by an Independent Researcher), and the
Actual Value Measured of Area Cleared of Plot
NOTE: n = 36.

correlates with lower rates of forest clearance (Godoy et al. 1997; Godoy and
Contreras 2001). The results of this study suggest that education may allow
people to estimate inputs more accurately and therefore gain greater effi-
ciency in the use of forests resources.
Relative to the measured area, estimations errors were larger in smaller
fields. As a consequence, people made larger errors when estimating the area
cleared from fallow forest, since plots from this forest type were smaller than
plots from old-growth forests. In contrast, people made accurate estimations
of the area of old-growth forest cleared because they typically cleared larger
plots from this forest type.
In sum, the analysis suggests that cross-sectional surveys will yield reli-
able estimates of area of old-growth and fallow forests cleared by indigenous
people provided one asks plot owners about the area of forest they cleared.
Asking household heads about the area cleared by the entire household will
yield inaccurate information.
REFERENCES
Bernard H. R., P. Killworth, D. Kronenfeld, and L. Sailer. 1984. The problem of informant accu-
racy. The validity of retrospective data. Annual Review of Anthropology 13:495–517.
Chicchón, A. 1992. Chimane resource use and market involvement in the Beni Biosphere
Reserve, Bolivia. Ph.D. diss., University of Florida.
Demmer, J., and H. Overman. 1999. Indigenous people conserving the rain forest? The effect of
wealth and markets on the economic behaviour of Tawahka Amerindians in Honduras.
Tropenbos series 19, Tropenbos International. Wageningen: the Netherlands.
Ellis, R. 1996. A taste for movement: An exploration of the social ethics of the Tsimane’ of Low-
land Bolivia. Ph.D. diss., St Andrew’s University, Scotland.
Finegan, B. 1996. Pattern and process in neotropical secondary rain forest. The first 100 years of
succession. Trends in Ecology and Evolution 11 (3): 119–24.
TABLE 2
Multiple Regression between the Absolute Value of Estimation Error by Plot Owner
(dependent variable) and Skills in Arithmetic and Education of Participant
Variable Coefficient Standard Error t p
Arithmetic skill 0.17 0.12 1.41 .17
Education –0.21 0.07 –2.77 .001
Constant 1.18 0.25 4.50 .0001
NOTE: Regression is ordinary least squares with robust standard errors; R2
= .19.

Godoy, R. 2001. Indians, markets, and rain forests. Theory, methods, analysis. New York:
Columbia University Press.
Godoy, R., and M. Contreras. 2001. Education, forest clearance, and environmental externalities
among lowland Indians in Bolivia. Economic Development and Cultural Change 49 (3):
555–74.
Godoy, R., P. Kostishack, K. O’Neill, and D. Wilkie. 1998. The socioeconomic correlates of
error in estimation of agricultural field size: An experimental study among the Tawahka
Indians of Honduras. Cultural Anthropology Methods Journal 10 (3): 48–53.
Godoy, R., K. O’Neill, S. Groff, P. Kostishack, A. Cubas, J. Demmer, K. MacSweeney, and
H. Overman. 1997. Household determinants of deforestation by Amerindians in Honduras.
World Development 25 (6): 977–87.
Grainger, A. 1993. Rates of deforestation in the humid tropics: Estimates and measurements.
The Geographical Journal 159 (1): 33–44.
Hecht, S. B. 1998. When solutions become drivers: The dynamics of deforestation in Bolivia.
Unpublished manuscript. Los Angeles: University of California, Department of Urban
Studies.
Huanca, T. 1999. Tsimane’ indigenous knowledge. Swidden fallow management and conserva-
tion. Ph.D. diss., University of Florida.
Lanly, J. P. 1982. Tropical forest resources. Food and Agricultural Administration forestry
paper no. 30. Rome, Italy: United Nations Food and Agriculture Organization.
Mathews, E. 2000. Understanding the FRA 2000. Forest briefing no. 1. Washington, DC: World
Resources Institute.
Myers,N.1992. Theprimarysource:Tropicalforestsandoutfuture.NewYork:W.W.Norton.
Nepstad, D., G. Carvalho, A. C. Barros, and P. Moutinho. 2001. Road paving, fire regime feed-
backs, and the future of the Amazon forest. Forest Ecology and Management 154 (3): 395–
407.
Picchi, D. 1991. The impact of an industrial agricultural project on the Bakairi Indians of Central
Brazil. Human Organization 50 (1): 26–38.
Pimentel, D., and G. H. Heichel. 1991. Energy efficiency and sustainability of farming systems.
In Soil management for sustainability, edited by R. Lal and J. F. Pierce, 113–23. Ankeny, IA:
Soil and Water Conservation.
Reyes-García, V. 2001. Indigenous people, ethnobotanical knowledge, and market economy. A
case study of the Tsimane’ Amerindians in Lowland Bolivia. Ph.D. diss., University of
Florida.
Romney, A. K., S. C. Weller, and W. H. Batchelder. 1986. Culture as consensus and informant
accuracy. American Anthropologist 88 (2): 313–38.
Silver, W. L., S. Brown, and A. E. Lugo. 1996. Effects of changes in biodiversity on ecosystem
function in tropical forest. Conservation Biology 10 (1): 17–24.
Simmons, C. 1997. Forest management practices of the Bayano region of Panama: Cultural vari-
ations. World Development 25 (6): 989–1000.
Smith, J., P. Van de Kop, K. Reategui, I. Lombardi, C. Sabogal, and A. Diaz. 1999. Dynamics of
secondary forest in slash and burn farming: Interactions among land use types in the Peru-
vian Amazon. Agriculture, Ecosystems and Environment 76:85–98.
United Nations Food and Agricultural Organization. 1993. Forest resource assessment 1990.
FAO forestry paper no. 112. Rome, Italy: United Nations Food and Agriculture
Organization.
302 FIELD METHODS

VINCENT VADEZ is a research associate from the Sustainable International Develop-
ment Program at the Heller School for Social Policy, Brandeis University. He is an
agronomist and plant physiologist who has been working with legumes since 1990. Since
1999, he has taken a turn toward the study of subsistence farming. His recent publica-
tions include “Does Integration to the Market Homogenize Agriculture? Evidence from
Tsimane’ Amerindians” (Human Ecology, 2003) and “Is Ethnobotanical Held Commu-
nally? Evidence from Bolivian Amerindians” (with Victoria Reyes-García et al., Sci-
ence, 2003).
VICTORIA REYES-GARCÍA is a research associate at the Tropical Conservation and
Development Center at the University of Florida. She is currently using network analysis
to evaluate the reasons for adoption of new farm technologies. Some of her recent publi-
cations are “Correlates of Delay-Discount Rates: Evidence from Tsimane’ Amerindians
of the Bolivian Rain Forest” (Journal of Economic Psychology, 2002), “Is
Ethnobotanical Knowledge Held Communally? Evidence from Bolivian Amerindians”
(with Vincent Vadez et al., Science, 2003), and “Measuring Culture as Shared Knowl-
edge: Do Data Collection Formats Matter? Cultural Knowledge of Plant Uses among
the Tsimane’ Amerindians of Bolivia” (Field Methods, 2003).
RICARDO GODOY is a professor in the Sustainable International Development Pro-
gram at the Heller School for Social Policy, Brandeis University. He holds master’s
degrees in social sciences and public administration, and he earned his Ph.D. in anthro-
pology at Columbia University. His main area of interest focuses on the effect of market
openness on human welfare and conservation among indigenous populations and the
use of experimental research design to assess high-profile public policy interventions.
Some of his recent publications are “The Life-Cycle, Economic, and Ecological Deter-
minants of Spousal Leisure Sharing: Panel Estimations from an Amerindian Society”
(Human Ecology, 2002) and “Local Financial Benefits of Rain Forests: Comparative
Evidence from Amerindian Societies in Bolivia and Honduras” (Ecological Economics,
2002).
LUKE WILLIAMS is currently earning a bachelor’s degree at the University of Florida
in mechanical engineering. He is employed by the physics department, where he designs
scientific equipment. His research interests are on gravity wave observation with laser
interferometery, solar energy conversion with Rankine cycle, and unmanned aerial
vehicles.
LILIAN APAZA has a bachelor of science degree in biology from the Universidad
Nacional Mayor de San Andrés, La Paz, Bolivia. She did her undergraduate thesis on
traditional uses of animals and animal medicines among the Tsimane’ and is currently
working on household surveys. Two recent publications are “Role of Meat Prices in
Household Consumption of Bushmeat among the Tsimane’ Amerindians of Bolivia”
(Oryx, 2002) and “Markets and the Use of Wild Animals for Traditional Medicine: A
Case Study among the Tsimane’ Amerindians of the Bolivian Rain Forest” (Journal of
Ethnobiology, 2002).
ELIZABETH BYRON is currently a doctoral candidate in the Department of Anthropol-
ogy at the University of Florida. Her doctoral research examines the relationship

between market integration and the physical health, nutritional status, health care utili-
zation, and dietary intake of lowland Amerindians in Bolivia. She received her M.A. in
anthropology at Florida in 1999. Her master’s thesis (1999), titled “Intra-household
and Community-Level Impacts of Participation in a Women’s Income-Generating Pro-
ject: Manabí, Ecuador,” analyzed elements of intrahousehold resource allocation and
economic and social empowerment. Her career interests include working in interna-
tional development in the areas of health, poverty alleviation, and gender equity.
TOMÁS HUANCA is a Bolivian scholar with a background of Aymara culture. At the
University of La Paz and the University of Oruro, he conducted research in the lowlands
with the Tsimane’ indigenous people. In 1999, he obtained his Ph.D. from the Depart-
ment of Anthropology of the University of Florida. He did his fieldwork among the
Tsimane’ and conducted research on their traditional swidden fallow agriculture sys-
tem. He trained a Tsimane’ to collect, transcribe, and edit the Tsimane’ oral history,
which has since been published. He is now receiving funding from the Dutch Coopera-
tion Agency to carry on ethnographic work on oral history and patronymic origin of sites
and territory. With the World Bank project, he has embraced a more scientific approach
in his work with indigenous culture.
WILLIAM R. LEONARD is a professor of anthropology at Northwestern University. His
research interests include biological anthropology, human ecology, human growth, and
nutrition. Some of his recent publications are “Food for Thought: Dietary Change Was a
Driving Force in Human Evolution” (Scientific American, 2002), “Energetics and Evo-
lution: An Emerging Research Domain) (with S. J. Zek, American Journal of Human
Biology, 2002), and “Ethnobotanical Knowledge Is Shared Widely among Tsimane’
Amerindians, Bolivia” (with V. Reyes-García et al., Science, 2003).
EDDY PÉREZ holds a B.S. degree in biology from the Universidad Mayor de San
Andrés, La Paz, Bolivia. He is currently working on a project on wildlife management at
the Estación Biológica of Tunquini in Bolivia. Some of his publications are “The Role of
Meat Prices in Household Consumption of Bushmeat among the Tsimane’ Amerindians
of Bolivia” (with several authors, Oryx, 2002) and “Daños Provocados por Fauna
Silvestre a Cultivos Agricolas Anuales en la Comunidad Chairo, Parque Nacional y
Area Natural de Manejo Integrado Cotapata-Bolivia (Ecología en Bolivia, in press).
DAVID WILKIE is a social scientist in the Living Landscapes Program of the Wildlife
Conservation Society. He has more than eighteen years of research experience in the
socioeconomic aspects of household-level natural resource use in Central and West
Africa and in Central and South America. Some of his research interests include examin-
ing the impacts of trade and the commercialization of nontimber forest products on for-
est animal populations, the role that logging plays in promoting bushmeat markets, the
income and price elasticities of demand for bushmeat, and the use of satellite imagery
and aerial photography to model the location, extent, and rate of land transformation
within rain forests. He has recently published “Economics of Bushmeat Consumption”
(Science, 2000).
304 FIELD METHODS

Measure2003

More Related Content

What's hot (20)

Similar to Measure2003 (20)

More from Lilian Apaza (17)

Recently uploaded (20)

Measure2003