Donald Green and Alan Gerber from Yale University wanted to find out if Registration Based Sampling provided more accurate election result forecasts than did polling with the normal Random Digital Dialing system employed by many pollsters.
The answer, was yes, and that answer was the reason Wilson Perkins Allen Opinion Research has employed Registration Based Sampling for it’s clients.
The differences in the two systems are plenty, below I will describe a few:
1) A pollster, when using Random Digital Dialing, usually has to rely on the respondent to indicate whether they are registered to vote or not. Whereas with RBS, the list is generated already knowing the name of the respondent and whether or not they are registered and how they are registered.
2) But frequency of a voter is also important. When polling, a firm wants to be sure we are talking to voters who actually plan on voting and have a record of voting. In RDD, one is again relying on the respondent to be honest and not misstate their intention to vote leaving guesswork involved. In RBS, because the voter list is being used, frequency of a voter is included in the list and this removes any guesswork on behalf of the pollster.
3) The main reason for using RBS is the wealth of information available to the pollster that can be used to forecast voter turnout based on elections similar to those that are being polled at the time. A pollster can stratify the sample to base the demographics of the poll based on frequency to vote, age, gender, geographic region, party registration, etc. RDD simply does not provide for this type of accuracy.
4) In RDD pollsters must discard or weight down the interviews they conduct with those they deem to be unlikely to vote. This is a waste of resources to interview unlikely voters from the start, but RDD leaves the pollster with no option. RBS however gives the pollster the means to differentiate between likely and non likely voters BEFORE any resources are spent.
While there is so much more that can be said about the accuracy and reasons for using RBS compared to RDD, the above four reasons are solid reasons to make sure as a client you are getting the best service and accurate results that you are paying for.
It is accuracy of results, top-notch service for our clients and new and innovative research methods that WPA continues to strive for on a daily basis, and one reason we always use Green and Gerber’s Registration Based Sampling method in all of our survey research
