How information about job openings is distributed is held to be a central d
eterminant of social stratification. Interest here has focused on social ne
tworks, yet methodological problems in earlier research prevent definite co
nclusions from being drawn. Equally serious are the limitations associated
with the setting of earlier research, since previous studies have examined
the impact of networks in situations in which contact networks would be exp
ected to matter. This article instead explores the limits of social network
s by highlighting a situation in which networks are less likely to be of gr
eat importance: job search among the unemployed in Sweden. Contact networks
have been assumed to be less useful for unemployed job searchers, and Swed
en has a substantively important alternative information channel in the pub
lic employment agencies. Evidence is offered here on how the employment pro
bability is affected by the use of contacts during the period of job search
, with the problem of selection bias taken into explicit consideration. The
results indicate network size to be positively related to the probability
of employment, but question the importance of tie strength.