A new report from the Pew Center investigates this important question. Polls can be highly influential–a not insignificant number of people want to be part of a ‘winning’ team and so use the polls to decide who the ‘winner’ is likely to be, and add their vote to that group. But is there a difference between a poll like the Harris poll and a poll like the one Andy Cohen* does on Watch What Happens Live (A Bravo-TV talk show highlight Bravo Reality show ‘stars’?)? Why yes, yes there is a difference. And that’s what the Pew Center article talks about (although they don’t mention Andy Cohen’s poll in particular).
The report talks about the challenges methodologically. For example:
- a polling company that uses only landline phone numbers will miss half the population
- using cell phone numbers to sample from is better, but many people move to a different state and keep their phone number. If pollsters use area codes to project responses by state, then that information is off: 10% of adults have a cell phone number that doesn’t match where they live
- election polls are tricky because pollsters need to talk to ‘likely’ voters: the 60% of the population that actually votes.
- Online polls are not random samples, and are also problematic.
It is an interesting report if you ever want to challenge the results of a poll!
Credits for the report Flashpoints in Polling:
*Random fact: Andy Cohen and I went to the same high school, Clayton High School in Clayton, MO. Go Greyhounds!
MTurk for Academics 