There are many different ways that samples can be obtained for online surveys.  These include open invitation surveys of volunteers, intercept surveys, opt-in or access panels, Amazon’s Mechanical Turk, Google Consumer Surveys, list-based samples, and the like.  In most cases, the goal is to make inference to some large population.  The different approaches to selecting samples and inviting respondents to complete a survey vary in their inferential properties.  Threats to inference include sampling error, coverage error, and nonresponse error.  In addition to selection methods, a variety of adjustment methods, such as weighting, propensity score adjustment and matching, are being used to mitigate the risk of inferential errors.  The course will focus on the assumptions behind the different approaches to inference in Web surveys, the benefits and risks inherent in the different approaches, and the appropriate use of a particular approach to sample selection in Web surveys.  The course has a conceptual rather than statistical focus, but a basic understanding of statistics will be helpful.
Course Homepage