This recent study highlighted by FiveThirtyEight may show the way to get around a huge problem in public surveys. With falling numbers of land-line phones, call screening and being overwhelmed by telemarketers, survey firms are increasingly unable to gather representative data from targeted populations. This new technique offers a two-stage method that might get higher response rates. First, use an on-line survey to identify those who might be willing to accept survey calls. That survey could be used to find the baseline characteristics of the survey population. Then second is to make the calls with targeted and/or in-depth questions. This could be a promising answer to an ongoing problem.
Literary Digest, anyone?
Literary Digest (conducted in 1936 predicting a loss by FDR for those wondering about the reference) was a single-shot voluntary survey. Recent Internet surveys have similar problems although they correct for some biases. This new method would correct for the bias problem in the Literary Digest approach.