The Commonwealth Fund released a report in December on young people’s views of health insurance. Most of the results were relatively unsurprising, however the survey methodology would make any statistician excited.
Traditionally, public opinion surveys have been conducted over the telephone. However, young people typically don’t have landline telephones. Many, if not most, use cell phones primarily or exclusively. 83% of all adults have cell phones and for a vast majority of people this is becoming their main form of communication.
Data from young people is valuable to public opinion researchers. Yet, with caller identification technology so readily available on cell phones, individuals are less likely to answer calls from researchers. This creates an interesting problem– now researchers must find out exactly how many people primarily use cell phones, and the traditional way to ask them is by calling them. So how do researchers conduct a survey today?
In this survey, about half of the participants were called on their cell phones and the other half was called on landlines. The second group of landline calls were over-sampled for low income and African-American and Hispanic populations. Is this really accurate to do? How do we know the best way to sample? Are our methods for collecting data truly reflecting our population? This is why trusting statistics blindly does not often yield valid results, as we are seeing with the recent Massachusetts Senate race polling. Methodology is more important than the results of a given survey.
If anyone was wondering, the Commonwealth survey found that almost half of young people had gaps in insurance and low income individuals were more likely to be uninsured.