We describe work in which we used three days of interviewing to identi
fy and measure the network among 200 people significant in a complex p
roduction process. The capture-recapture strategy should be useful in
other settings: (1) Conduct survey network interviews with people (inf
ormants) positioned in the study population such that their contacts o
verlap to provide recaptured relations. (2) Estimate reliability from
data consistency across recaptures. (3) Triangulate relation response
categories to assign quantitative scores to the categories. (4) Use re
liability correlates to weight recaptured relations in the final netwo
rk pooled across interviews. (5) Extrapolate from the known strengths
of the captured relations to define the uncaptured relations.