Web crawling, snowball sampling, and respondent-driven sampling (RDS) are three types of network driven sampling techniques that are popular when it is difficult to contact individuals in the population of interest. This talk will show that if participants refer too many other participants, then under the standard Markov model in the RDS literature, the standard approaches do not provide "square root n" consistent estimators. In fact, there is a critical threshold where the design effect of network sampling grows with the sample size. This critical threshold depends on the spectral properties of the underlying social network. To ensure that our estimates remain “square root n” consistent, we will discuss two novel approaches. The first approach must be incorporated during data collection. The second approach is a novel estimator.