The Fit Between Theory and Data in Respondent-Driven Sampling: Response to Heimer