Don’t be thrown by a recent flurry of New York State polls on the governor’s contest between “The Son also Rises” Andrew Cuomo and “I’ll clean up Albany with a baseball bat” Carl Paladino. Is Cuomo ahead by 6 points (Quinnipiac) or 33 points (Siena), or somewhere in between, 19 points (Marist)? Much of the difference can be explained in the varied methodologies of the polling organizations. Were the numbers based upon registered or likely voters? Was former Conservative Party candidate Rick Lazio included in the tossup question?
Public opinion polling has a statistical basis but a great deal of the sausage making has to do with the judgments and interpretations of the various pollsters. How are respondents selected? Are cell phones being used as well as landlines? What effort is being made to reach hard to reach voters? What is the question wording and order? Is the quality of the interviewing up to industry standards? How is the data balanced to ensure it reflects the electorate? AND, so much more…
Clearly, not all polls are created equally. Given the range of possibilities each poll organization can utilize, it may be more surprising that poll results are often similar and not all over the map as recently occurred in the New York governor’s race.
Now, this electorate is a tougher read than in recent election cycles, especially when it comes to who is likely to turn out. It is a volatile time, and the polls will need to pick up on that uncertainty. Having said that, I suspect the next round of Cuomo-Paladino polls to be singing a similar tune. It is also unlikely the media will serve up a similar amount of coverage to the “polls are similar” story as the recent avalanche of words devoted to the “why polls are so different.”