Methods and data
3.1. Data description
To assess the roles of the six factors on participation, and in turn
the role of participation on institutional outcomes, we analyze data
from two different sources. The first source is a dataset of
household surveys administered in 23 sites in four countries
(Bolivia, Kenya, Mexico, and Uganda) identified by the International
Forestry Resources and Institutions (IFRI) program in 2008
(see Wollenberg et al., 2007 for a discussion of IFRI research
protocols.). The second source is a more aggregated dataset taken
from 96 forest associations in 10 countries of the IFRI program.
Table 1 reports the salient features of both data sources as it applies
to the models presented in Equations (1) and (2).
The two data sources have different strengths and weaknesses.
The household data has a much larger sample size, but is limited in
that participation is measured only for the person in the household
with the primary responsibility to attend FUG meetings. In
addition, there is no data on entrenched claims as an explanatory
variable. The IFRI forest association data is limited in that there are
much fewer observations and less detail about how households
individually participate. Analysis at the association level does
permit us to capture the effects on the whole community from
greater women’s participation, but the level of aggregation does
not permit us to see how individual households respond when
women from their household participate. Using both data sets
provides a clearer picture of the determinants and effects of
gendered participation at multiple levels.