We dichotomized (converted to binary data) each stakeholder
group’s Pearson’s correlation matrix in UCINET version 6.365
(Borgatti et al., 2002). The cutoff points were selected to include
the coefficients representing 40% of the strongest negative, and40% of the strongest positive correlations for each stakeholder
group. Because the sample size of fishers, managers, and scientists
differed, we decided to use a cut off point that represents a
consistent proportion of correlations rather than a significance
level which would be biased by sample size. The network analysis
therefore gives an indication of the tendencies towards a trade-off
or synergy within groups rather than an indication of the strength
or number of connections across groups. It is worth noting that the
manager network in particular is composed of a fairly small sample
size, so although the positive and negative trends should be fairly
robust, they will inherently have greater uncertainty associated
with them. We produced two matrices for each stakeholder, one
representing trade-offs (negative) and one representing synergies
(positive). For each stakeholder group, we combined the positive
and negative matrices in NetDraw (Borgatti, 2002) to create a
network representing stakeholders’ synergies and tradeoffs in
ecosystem service priorities. For each network, we calculated two
types of centrality measures, degree and betweenness. Degree
centrality is a measure of the number of connections between a
service and all other services. Betweenness centrality is a measure
of the number of shortest paths that run through a service,
representing the importance of a service for connecting other
services that would otherwise be unconnected (Wasserman and
Faust, 1994).