A simulative comparison of respondent driven sampling with incentivized snowball sampling--the "strudel effect".