Coverage Error is due to the fact that not every unit in the
population is represented on the frame. For example persons without Internet access
are by definition excluded from the target population of a web survey.
However, web surveys and online polls tend to advertise their results as
representative of general public
opinion. See box on the left for proportion of Americans that have access
to the internet.This illustrates that coverage error takes into account the opinions of individuals who have graduate education and who are between the ages of 18 – 29. Web surveys are flawed by virtue of the fact that there is a digital divide in our nation.
Even more worrying is the practice of RDD
(Random Digit Dialing) surveys which survey
opinions of random individuals who own fixed phone lines. This is a common
practice among surveyors to introduce a level of accuracy in terms of
population that is being sampled. It is thought that RDD
surveys give a more accurate picture of public
opinion. However, this practice is in great danger of becoming irrelevant
as RDD
sampling does not take into account portion of the population that does not
have telephones. This is even more pertinent today, especially with the 18 – 24
yr old category of the population, who tend to rely solely on cell phones and
do not own fixed line telephones. Surveyors are not able to perform a random
dial survey on cell phone numbers because the Federal Communications Commission
also has erected do-not-call rules barring survey researchers from using RDD,
to find cell-phone users. These rules also prohibit calling people who have to
pay for the call.
Next : Measurement Errors
Previous : Poll-O-Rama: Actual Poll Examples from the Media