A CNN poll released November 21 has Newt Gingrich leading Mitt Romney 24 percent to 20 percent, putting Gingrich in the lead for the GOP nomination. Although candidate Herman Cain took a hit at the polls following harassment allegations and several public speaking gaffes, this poll indicates that he may be recovering, as he polled in third place at 17 percent. Although this poll suggests a dramatic change from previous polling, respondents indicated they still have not made up their minds, as only 27 percent answered that they would definitely support the candidate they selected in the 2012 election, which is down from the 31 percent that declared definite support in the previous two CNN polls which were conducted in mid-November and mid-October.
The CNN poll consisted of 1019 adult Americans contacted between November 18 and November 20. Respondents were contacted via random dialing of both landlines and cell phones by live interviewers. Breakdowns of telephone contacts included 818 landline interviews and 201 cell phone interviews. Questions pertaining to the Republican primary were only delivered to 402 Republican voters within the sample, however CNN has not released whether “Republican voters” were defined based on voter registration or stated willingness to vote in the GOP primary. The margin of error for the general sample was +/-3 points, and +/- 5 points for the smaller sample of Republican voters.
Was the Poll Biased?
The poll methodology is mostly in line with NCPP standards and the scientific method. There is some information that was left out of CNN’s poll release that bring up questions about potential biases that may or may not have been present. One of these is the way in which likely Republican voters were identified from within the general sample. The Republican respondents will be more likely to reflect actual primary voters if they were asked whether they plan to vote in the primaries, as opposed to being asked whether they are registered Republicans, as many registered voters forgo primary elections. CNN also does not reveal any information about weighting of the Republican sample. Weighting likely occurred with such a small group of respondents, but readers of the poll are not informed as the the extent or direction of the weighting.
Additionally, the poll questions released by CNN start at question number 27. Presumably, the previous questions were directed at the general sample, and the release declaring Gingrich the new frontrunner only applied to questions asked of the small Republican sample, but questions previously asked to the respondents may have affected subsequent answers.
Interpreting the Results
Poll aggregators such as RealClearPolitics are valuable resources when analyzing the validity of individual polls. Scientific theories are created when patterns of experiment results trend together to form a conclusion, so, polling results are more scientific when many polls yield the same result. This appears to be the case with Gingrich’s lead, as RealClearPolitics currently has him as the GOP frontrunner, with a 2 point lead over Romney, based on averaged results from several recent polls which include this particular CNN poll.
CNN/ORC International Poll. Accessed November 22, 2011
RealClearPolitics. 2012 Republican Presidential Nomination. Accessed November 22, 2011.
Decoding Science. One article at a time.