Two recent polls seem to be at odds concerning President Obama’s approval rating: which is more accurate – Gallup, which shows Obama with a more positive approval rating, or Rasmussen, which shows Presidential Approval underwater?
Gallup daily tracking, conducted from April 20 to April 22, has Obama’s approval on the uptick – with 50 percent of respondents approving and 44 percent disapproving. Rasmussen’s daily tracking poll, conducted over the same period of time, is nearly a complete opposite, with 45 percent approving and 54 percent disapproving.
Why are these numbers so different?
Obama Poll: Question Wording Matters!
Gallup asked its respondents to select “approve” or “disapprove.” Rasmussen offered some degree of options, with those who selected “strongly approve” and “somewhat approve” being lumped into the “approve” category, and those who selected “strongly disapprove” and “somewhat disapprove” being lumped into the “disapprove” category. Which of these methods leads to a more accurate result is debatable. On one hand, voters who “lean” in one direction may be more comfortable expressing their choice with a “somewhat” in front of it, and less likely to refrain from the question all together. On the other hand, voters may use the “somewhat” as a good reason to embrace their impulsiveness, as in “I’m not a fan of Obama, but my brother just got a new job, so I guess I somewhat approve,” or “I really like Obama, but my favorite stock went down a little bit today, so I guess I somewhat disapprove.”
If these voters stand a good chance of feeling differently toward Obama a few months from now, it is questionable whether their response should be considered the same as solid approvals and disapprovals, or if they should be considered undecided. Rasmussen does address this concern by reporting a Presidential Approval Index based only on the “strongly approve” responses, however this index is not considered when the polling agency reports its overall presidential approval numbers.
Likely Voters Versus All Americans
Rasmussen compiles results from respondents who state that they are likely to vote in the 2012 presidential election, while Gallup conducts a random survey of Americans. Some analysts consider Rasmussen’s model to be more accurate, since it weeds out respondents who have no intention of voting, and may not be politically saavy enough decide whether or not they approve of the president. Those who are more supportive of Gallup’s method argue that probability should resolve the issue of politically inactive respondents, as long as the sample is large and diverse, and weighting is used appropriately.
Decoding Science. One article at a time.