There is a classic conflict between accessing low incidence populations and the principle of randomness.
A client may want to interview highly affluent individuals or those with a medical condition that affects <1% of the population. How can that be done in a way that is both cost effective and gives you the confidence that the sample of respondents accurately reflects the low incidence population?
Randomly screening for these individuals would guarantee that perhaps 99% of those whom you interview would not be eligible (screen out). The cost for a project like this would be prohibitive.
Acquiring lists of these difficult to reach groups would lower the cost of data collection, but raised the question of how representative they are. How like or unlike the greater population of those with the condition or characteristic are they?
RDR understands the conflict between our clients' need for information collected in a cost effective manner and the risk of providing data that are flawed. These are the details that we, as your research partner, need to "sweat." Flawed data is much worse than no data.
That is why we continually compare our data to appropriate third party sources. Often we access the only organization that can afford to conduct random surveys, especially with low incidence populations: the federal government and organizations they fund. We scour these data sources for incidence and profiles of these rare populations.
We are then able to compare the data we collect and even quota sample so that our respondents look as much like the population as possible. Upon the conclusion of our data collection, we carefully weight those respondents to look like the population.
While not the perfect solution it is an optimal solution of the balance between accuracy and cost effectiveness.
