Information from social networks (e.g. its population structure) can be a source of auxiliary information for improving the precision of estimators of the population total or mean relative to methods based on the nonparametric Horvitz-Thompson estimator.

Results are based on simulation and show that:
1.incorporating network information via a contextual model seems to be the most appropriate approach;

2. only knowledge of the partial network linking the sampled population units to the non-sampled population units is necessary.

An estimator for the mean-squared error to make an informed decision about using the contextual information is also introduced. Results show that this adaptive strategy leads to higher precision.