Abstract This article presents the results of a pilot feasibility study comparing two alternative recruitment approaches based on Respondent Driven Sampling using initial seeds selected from a US nationally representative panel, AmeriSpeak, to augment the number of lesbian, gay, bisexual, and transgender (LGBT) respondents to a short web survey on smoking, discrimination, and health. In the nomination condition after completing the survey both LGBT and non-LGBT seeds were invited to share the names and email address of up to four LGBT persons they knew. In the recruitment condition, seeds were given four unique PINs and links to the survey to distribute to LGBT persons. Both conditions were successful in producing new LGBT respondents. The recruitment condition was much more productive. LGBT seeds (and their recruits) were much connected to and willing to contact other LGBT people they knew to participate in a survey. Comparisons of characteristics and responses from the initial samples and the LGBT referrals as well as comparisons to LGB samples from a large national survey are presented. Results demonstrate the promise of this hybrid technique for increasing the number of LGBT respondents through referrals from an initial probability based sample.

中文翻译：

---

对LGBT种群进行抽样的混合方法的测试：来自概率样本的种子的Web响应者驱动抽样

摘要本文介绍了一项试点可行性研究的结果，该研究比较了基于响应驱动抽样的两种替代招募方法，该方法使用了美国国家代表小组AmeriSpeak选出的初始种子，以增加女同性恋，男同性恋，双性恋和变性者（LGBT）的人数有关吸烟，歧视和健康的简短网络调查的受访者。在完成调查后的提名条件下，LGBT和非LGBT种子均被邀请分享他们认识的最多四个LGBT人员的姓名和电子邮件地址。在招聘条件下，为种子提供了四个唯一的PIN，并提供了与调查的链接以分发给LGBT人员。两种情况都成功地吸引了新的LGBT受访者。招聘条件更具生产力。LGBT种子（和他们的新兵）与他们知道要参加调查的其他LGBT人员紧密联系，并愿意与其联系。介绍了原始样本和LGBT推荐人的特征和响应的比较，以及来自大型国家调查的LGB样本的比较。结果表明，这种混合技术有望通过从基于初始概率的样本中进行推荐来增加LGBT受访者的数量。